Master Art of Translation | The Reported Speech Boundary Fidelity System — How Direct Quotes, Indirect Reports and Reporting Verbs Preserve Speaker Responsibility

Accurate reported speech translation means preserving who said what, to whom, when, with what degree of commitment, and whether a sentence is a direct quotation, an indirect report, a paraphrase, or a narrator’s interpretation. People searching for how to translate reported speech, direct and indirect speech translation, say versus tell, reporting verbs, quotation marks in translation, or the difference between a quote and a paraphrase need more than a tense-conversion chart. They need a way to stop another person’s words from quietly becoming the translator’s assertion.

The hidden problem is responsibility. A sentence may contain excellent vocabulary and fluent grammar while transferring a claim from a witness to a journalist, changing a warning into an order, converting an uncertain recollection into a verbatim quote, or making a fictional narrator sound as though a character spoke aloud. Even the choice between ‘said’, ‘claimed’, ‘confirmed’, ‘admitted’, ‘denied’, ‘suggested’ and ‘promised’ can change what readers think happened. Translation must preserve the boundaries around speech events before it can safely polish their wording.

This guide gives a practical system: recover the original speech event; identify the speaker, audience and quoted span; distinguish direct words from indirect content; preserve the force of reporting verbs; adjust tense, person, time and place only where the target perspective requires it; then run independent-reader checks against false attribution, added certainty and lost evidence. You will find a repeatable method, failure-mode diagnostics, original worked cases and exercises that transfer across languages, professions, fiction and ordinary conversation. The examples use English as a transparent analytical gloss, not as a claim that every language follows English grammar.

• Search intent: translate direct and indirect speech accurately; avoid reported-speech translation mistakes; select reporting verbs without inventing force.
• Hidden problem: the boundary between someone’s words, someone’s account of those words, and the translator’s own interpretation.
• Mechanism: map speaker, utterance, listener, report author, time, evidence and speech act before choosing words.
• Method: source-event reconstruction → quote classification → attribution control → target-language reconstruction → validation.
• Diagnostics: verify quoted versus paraphrased material, reported questions, commands, denials, corrections and speaker-switches.
• Transfer: reporting in literature, research, journalism, meetings, customer service, institutional notices and multilingual records.
• Next routes: return to the translation master root; use the existing Vocabulary and How English Works owners only for their respective learning jobs.

1. The real object being translated: a speech event, not an isolated sentence

When the source reads ‘She said the system was safe,’ the grammatical sentence contains at least two propositions: a person made an utterance and the utterance represented a system as safe. These propositions are not interchangeable. A target sentence that simply says ‘The system was safe’ deletes the attribution and may increase the writer’s apparent responsibility for the claim. A target that says ‘She proved the system was safe’ adds a verification event that the source did not describe. The task is to preserve both the content inside the report and the frame that tells readers how that content entered the text.

A useful source-event record contains separate slots for the original speaker, the original listener, the quoted or reported content, the communicative action, the reporter, the reporting moment, the speaker’s evidence, and any editorial interventions. Not every source will fill every slot. Blank does not mean permission to invent. If no original listener is specified, the translation should not casually introduce one. If the source merely says ‘reportedly’, the translation should not manufacture a named source. If quotation marks are absent, do not assume an indirect summary is verbatim.

2. What direct quotation, indirect reporting and paraphrase actually preserve

Direct quotation presents words as belonging to a speaker in a form that invites exact-word accountability, although editing and quotation conventions still vary by genre. Indirect reporting transmits the content or action of an utterance without promising identical wording. Paraphrase is a form of re-expression and may summarise, compress or reorganise the speaker’s point. A narrated thought is not necessarily a spoken sentence, and a headline that compresses a claim is not automatically a faithful substitute for a speaker’s original words. Translation must keep these statuses visible to readers.

Consider three analytical glosses: ‘Mara said, “The valve is open.”’ ‘Mara said the valve was open.’ ‘The report attributes an open valve to Mara’s account.’ Each gives different access to wording, tense and source responsibility. They can sometimes communicate compatible facts, but they are not interchangeable records. In a novel, the exact voice may matter. In a court transcript, changing a quote into a paraphrase may be prohibited by the job. In a general article, an authorised summary may be appropriate but must remain signalled as such.

3. A compact decision table for the reporting boundary

• Direct words: protect the quoted span and the speaker label; change grammar only as the translation requires, never silently claim greater verbatim certainty.
• Indirect content: preserve reported proposition, speaker and communicative force; choose natural reporting constructions in the target language.
• Reported question: distinguish asking from asserting; preserve whether the question is open, polar, rhetorical or a request.
• Reported command: distinguish instruction, permission, request, advice and threat; preserve who is expected to act.
• Paraphrase: flag summarisation and keep the speaker’s scope, uncertainty and exclusions.
• Hearsay chain: track the number and identity of intermediaries; do not present third-hand information as direct testimony.
• Free indirect style: preserve ambiguous blending of character viewpoint and narrator voice when that is the literary technique.
• Editorial bracket or ellipsis: follow the genre’s quote-handling convention rather than hiding translator interventions.

4. Twenty-eight mechanisms that change how a report should be translated

1. Verbatim status

A quote invites the reader to attribute a particular string of words to a speaker. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Decide whether the source offers precise words, remembered words or merely a content summary. Write this decision in the margin as a short statement of the invariant. Diagnostic: If the target is recited as though from a recording, would the original evidence support that claim? Typical false improvement: An ordinary paraphrase converted into quotation marks creates an invented transcript. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

2. Reporting clause attachment

A speech tag may apply to one clause, several sentences or a whole narrative passage. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Mark the exact span governed by the reporting verb before rearranging word order. Write this decision in the margin as a short statement of the invariant. Diagnostic: Can a reviewer draw the same boundary around reported material in both versions? Typical false improvement: Moving a tag across a sentence boundary can transfer responsibility for a judgement. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

3. Say versus tell

English distinguishes unmarked reporting with say from addressee-oriented tell; other languages have different patterns. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Recover whether an addressee is explicit, then use idiomatic target valency. Write this decision in the margin as a short statement of the invariant. Diagnostic: Does the target introduce an addressee absent from the source? Typical false improvement: A forced word-for-word mapping may create an ungrammatical or socially different account. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

4. Claim

Claim often places a proposition inside a speaker’s assertion without asking the narrator to endorse it. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Keep the assertion with the original claimant and resist strengthening it into fact. Write this decision in the margin as a short statement of the invariant. Diagnostic: Would a reader mistake the claimant’s position for independent verification? Typical false improvement: Replacing claimed with demonstrated converts source responsibility into narrator certainty. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

5. Confirm

Confirm ordinarily presents verification or ratification, not merely saying something. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Check whether the source really asserts confirmation and whether the object of confirmation is explicit. Write this decision in the margin as a short statement of the invariant. Diagnostic: What earlier statement, test or event was confirmed? Typical false improvement: A neutral speech verb translated as confirmed invents a stronger relation to evidence. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

6. Admit

Admit can position the speaker as conceding something against interest or after resistance. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Preserve confession or concession only if the source carries it; consider context. Write this decision in the margin as a short statement of the invariant. Diagnostic: Could the speaker reasonably deny that the remark was an admission? Typical false improvement: Turning ‘said’ into ‘admitted’ adds judgement and narrative history. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

7. Deny

A denial is a speech act about a proposition, not necessarily proof that the proposition is false. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Represent the exact proposition under denial and the scope of negation. Write this decision in the margin as a short statement of the invariant. Diagnostic: What was denied: the event, responsibility, timing, knowledge or motive? Typical false improvement: An ambiguous negated report can reverse the speaker’s intended response. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

8. Allege

An allegation signals a claim whose evidential or procedural standing matters. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Avoid words that imply conviction, proof or absolute truth when only allegations are reported. Write this decision in the margin as a short statement of the invariant. Diagnostic: Would a reader know the matter is attributed and not established? Typical false improvement: Unattributed accusations expose subjects and readers to misrepresentation. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

9. Suggest

Suggest can signal tentative inference, proposal, implication or advice. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Disambiguate from surrounding context before choosing one target construction. Write this decision in the margin as a short statement of the invariant. Diagnostic: Can the suggestion be accepted as advice, or is it a tentative finding? Typical false improvement: A suggestion translated as an order destroys its interpersonal force. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

10. Warn

A warning connects a possible danger, an addressee and an intended protective effect. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Keep the hazard, conditionality and the speaker’s warning action separate. Write this decision in the margin as a short statement of the invariant. Diagnostic: Was the speaker warning about danger or threatening to cause it? Typical false improvement: Confusing warning with threat changes who appears responsible for harm. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

11. Promise

A promise commits its speaker to future action, subject to the wording and situation. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Preserve speaker, beneficiary, event and conditions; do not elevate hope into promise. Write this decision in the margin as a short statement of the invariant. Diagnostic: Could the addressee rely on this statement as an undertaking? Typical false improvement: Turning ‘hopes to return’ into ‘promised to return’ invents an obligation. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

12. Order

An order invokes authority and compliance, not just a future statement. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Check the authority relation and whether the target verb imposes stronger force. Write this decision in the margin as a short statement of the invariant. Diagnostic: Did the utterance command, ask, recommend or authorise? Typical false improvement: A polite request mistranslated as an order can change institutional procedure. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

13. Request

A request asks someone to act without necessarily claiming the authority of an order. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Preserve the requested action, addressee, politeness and deadline. Write this decision in the margin as a short statement of the invariant. Diagnostic: Does the target make compliance compulsory when the source did not? Typical false improvement: A command-like target can misrepresent interpersonal hierarchy. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

14. Apologise

An apology is a speech act that may be conveyed with many possible words. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Translate the act when authorised, but do not infer an apology from a neutral expression of regret alone. Write this decision in the margin as a short statement of the invariant. Diagnostic: Does the source actually accept responsibility or merely acknowledge inconvenience? Typical false improvement: A fabricated apology can be materially consequential in disputes. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

15. Concede

Concession often repositions a speaker in an argument while retaining a counterpoint. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Track what is conceded and which contrary proposition remains defended. Write this decision in the margin as a short statement of the invariant. Diagnostic: What claim was surrendered, narrowed or accepted? Typical false improvement: Compressing a concession into total agreement reverses argumentative structure. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

16. Ask

Questions are speech acts, not simply statements with punctuation. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Keep open questions, yes/no questions and requests for information distinct. Write this decision in the margin as a short statement of the invariant. Diagnostic: Would the report allow the reader to know what answer was being sought? Typical false improvement: Reported questions may acquire the grammar of statements but should not become assertions. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

17. Wonder

Wonder frames an internal question or uncertainty and may not be addressed to anyone. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Recover whether the expression is spoken aloud, written or merely thought. Write this decision in the margin as a short statement of the invariant. Diagnostic: Who, if anyone, heard the question? Typical false improvement: Translating internal wondering as public asking invents a social event. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

18. Quote within a quote

Nested reported speech has at least two speakers and two frames of accountability. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Track layers with a stack rather than flattening everyone into one source. Write this decision in the margin as a short statement of the invariant. Diagnostic: Whose words are being quoted at each nesting level? Typical false improvement: A nested quotation attributed to the outer speaker can falsify a record. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

19. Reporting from documents

A statement in a document can be reported by an author, department or signatory. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Distinguish the person who wrote, the organisation responsible and the entity quoted. Write this decision in the margin as a short statement of the invariant. Diagnostic: Does ‘the report says’ misleadingly imply unanimous institutional endorsement? Typical false improvement: Source chains break when a document’s internal quote becomes the document’s conclusion. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

20. Backshift

Many English indirect reports permit changes in tense relative to reporting time. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Analyse event time and current relevance rather than applying a universal mechanical shift. Write this decision in the margin as a short statement of the invariant. Diagnostic: Is the reported event still current, finished or scheduled? Typical false improvement: Automatic tense conversion can move a future commitment into the past. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

21. Deictic recentering

Pronouns, now, here, today and tomorrow may depend on whose viewpoint is active. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Create a reference map for the original speaking moment and the later report. Write this decision in the margin as a short statement of the invariant. Diagnostic: Who is ‘I’; where is ‘here’; which date is ‘tomorrow’? Typical false improvement: Moving a quotation into indirect discourse without recalculation misdirects readers. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

22. Free indirect discourse

Narrator grammar may carry a character’s emotional or evaluative viewpoint. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Look for voice leakage, question-like thought and character-specific diction. Write this decision in the margin as a short statement of the invariant. Diagnostic: Is this narrator fact or character perception? Typical false improvement: Tidying the sentence into omniscient reporting destroys narrative ambiguity. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

23. Thought versus speech

Thinking, believing, remembering and saying are different event classes. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Do not turn private cognition into voiced assertion without clear source support. Write this decision in the margin as a short statement of the invariant. Diagnostic: Is the reported content audible, written, imagined or inferred? Typical false improvement: A character’s fear presented as public testimony changes the plot. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

24. Headline compression

Headlines frequently omit reporting auxiliaries and background attribution. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Recover the article’s source frame before translating a compressed headline. Write this decision in the margin as a short statement of the invariant. Diagnostic: What does the headline assert directly and what does it attribute? Typical false improvement: A catchy headline can silently turn a disputed claim into apparent fact. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

25. Partial quotation

A short quoted fragment can be embedded in a writer’s paraphrase. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Keep the quotation boundaries attached to the exact original fragment. Write this decision in the margin as a short statement of the invariant. Diagnostic: Which words are quoted and which are the journalist’s stitching? Typical false improvement: Moving quotation marks changes the perceived exact words. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

26. Translation of source-language quotes

A translated quote is not the same thing as an original-language transcript. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Follow editorial norms for signalling that a quote was translated when material. Write this decision in the margin as a short statement of the invariant. Diagnostic: Could readers mistake target phrasing for the source speaker’s original language? Typical false improvement: Unmarked translation can falsely imply the speaker chose particular target-language connotations. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

27. Correction in the record

A speaker may retract, repair or contradict an earlier statement. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Preserve sequence, changed proposition and who initiated the correction. Write this decision in the margin as a short statement of the invariant. Diagnostic: Does the final report preserve that a correction happened? Typical false improvement: Only translating the polished final statement deletes evidence of repair. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

28. Silence and non-answer

A refusal to answer, evasion or silence is not an affirmative statement. This matters because translation of reported speech reconstructs a relationship between an utterance and a later account of that utterance. The original form may not survive intact, but its responsibility structure must. Before selecting a reporting verb or rearranging clauses, identify the claim or action the source actually commits to and separate it from what a reader could merely infer. Otherwise fluency disguises a change in evidence or social force.

Operational choice: Represent the absence of a reported proposition accurately. Write this decision in the margin as a short statement of the invariant. Diagnostic: What was not answered and what did the speaker actually do? Typical false improvement: A summariser can wrongly invent a stance from a non-answer. A defensible target may require a different grammatical construction, but it cannot be justified solely because the target sounds more elegant. Compare source and target from the perspective of a reader who does not know the source-language wording.

5. The twenty-eight-gate repeatable method: from source record to safe target

Gate 1: Locate the reporting span

Mark exactly where a report begins and ends; note punctuation and paragraph breaks. A dialogue tag can govern more than one clause, but a later narrator remark may not be attributed to the character. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 2: Identify the original speaker

Record the named or unnamed person, institution, character or document credited with the words. If a pronoun is ambiguous, resolve it from context or retain justified ambiguity rather than guessing. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 3: Identify the report author

Distinguish the person telling readers about the utterance from the person who made it. The journalist, narrator, researcher and translator may each form a different layer of representation. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 4: Identify the original audience

Determine who was directly addressed and whether the audience is named. Say, tell, ask and warn may depend on the presence of an addressee, but do not invent one. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 5: Classify evidence status

Decide whether the source supplies recording, remembered quote, transcript, interview note or second-hand account. The translation cannot be more verbatim than the available source warrants. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 6: Classify mode

Mark speech, writing, internal thought, signed communication, public announcement or quoted document. The same content changes meaning if a private thought becomes a public promise. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 7: Classify speech act

Choose the event as assertion, question, command, warning, request, promise, denial, admission and so on. A reporting verb is a small theory of what occurred, not a synonym chosen for stylistic variety. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 8: Map propositions

Write the core content as minimal claims and negations. ‘Not all’ must not become ‘none’; ‘not by Tuesday’ must not become ‘never’. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 9: Map speaker commitment

Record whether the speaker endorses, doubts, attributes or merely mentions the proposition. The reporter and the speaker can have different commitments to the same proposition. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 10: Mark direct quotes

Draw a boundary around wording represented as directly uttered. Preserve the quoted status even when target-language punctuation conventions differ. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 11: Mark indirect reports

Identify passages that carry content rather than exact words. Do not add quotation marks around an elegant but reconstructed paraphrase. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 12: Mark nested frames

Number each level of quotation or hearsay. Keep ‘A said B reported C claimed’ from collapsing into ‘A confirmed’. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 13: Resolve pronouns

Create a compact mapping for I, you, we, she, they, their and reflexives. An apparent simple pronoun change may reverse responsibility for an action. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 14: Resolve time anchors

Identify speech date, reporting date and event date separately. Backshift and ‘tomorrow’ calculations are relative to actual anchor dates, not a single universal table. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 15: Resolve place anchors

Map here, there, this room and nearby to the original viewpoint. A target report may need an explicit place phrase when the source deictic no longer points correctly. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 16: Retain polarity

Check not, never, hardly, no longer, scarcely and negative replies. Reported denials and refusals are particularly vulnerable to polarity reversal. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 17: Retain conditions

Mark if, unless, provided that and exceptions inside the speech content. A promise subject to approval is not an unconditional promise. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 18: Check role valency

Choose target reporting grammar that correctly marks speaker and addressee. The mapping between speech verbs and argument structure is language-specific. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 19: Check register

Preserve whether the reported speech was formal, intimate, sarcastic, technical or cautious. Do not polish a nervous witness into a confident official speaker. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 20: Check directness

Determine whether words were blunt, indirect, euphemistic or implied. A target summary may state the act, but should not exaggerate deliberate intention. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 21: Check editorial interventions

Mark inserted explanatory brackets, elisions or reordered fragments. Do not let a clarification look like it came from the source speaker. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 22: Check multi-sentence continuity

Follow speaker changes across dialogue turns and paragraphs. An unattributed ‘she added’ can attach to the wrong participant after sentence reordering. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 23: Check documentary chain

Link each attributed statement to the correct document, meeting or record. A source quoted in a report is not automatically a conclusion endorsed by its editors. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 24: Check target idiom

Write natural target-language syntax after preserving the invariant. Literal English backshift or quote placement may be inappropriate in another language. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 25: Check genre contract

Verify quotation rules for fiction, research, journalism, legal transcripts or subtitles. The tolerance for paraphrase and the duty to preserve exact wording vary by genre and assignment. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 26: Back-translate responsibility

Have a reader restate who is responsible for each claim after seeing only the target. If they assign a claim to the narrator that belonged to a source, investigate. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 27: Review high-stakes differences

Escalate changes affecting liability, consent, accusation, diagnosis, safety, policy or testimony. A small verb substitution can have a large consequence when readers act on the report. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

Gate 28: Release with audit trail

Record unresolved questions, permitted editorial choices and final reviewer decisions. Revision should not erase the reasons a delicate reporting boundary was preserved. Treat this as a test with an observable outcome, not as advice to ‘be careful’. In a short text the observation may fit in a margin note; in a long assignment it belongs in a review sheet. State the source span, the proposed target relation, and the specific misunderstanding the test is designed to catch. If the answer cannot be verified from the supplied context, preserve uncertainty, ask for reference material when possible and flag the issue for review rather than converting it into a confident guess.

6. Worked diagnostics: recover the frame, then translate the report

The following cases are original teaching examples. ‘Source’ and ‘unsafe’ are English analytical glosses designed to expose the structural decision; they do not prescribe literal translations for every target language. Each case has five separate checks: the event record, the source-language boundary, the reporting-verb decision, perspective reconstruction, and the independent-reader result. Work through the checks before attempting polished target-language prose.

Case 1. Laboratory result: may versus confirmed

Source analytical gloss: The technician said, “The sample may be contaminated.” Unsafe rewrite: The technician confirmed the sample was contaminated. Defensible report: The technician reported possible contamination of the sample. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a provisional observation, carried by a technician in communication involving a laboratory team. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is may versus confirmed. Because a test result must not be upgraded into proof, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this laboratory result setting, a test result must not be upgraded into proof. Retain uncertainty before a follow-up analysis is completed. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a technician and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is may versus confirmed. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Retain uncertainty before a follow-up analysis is completed. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the laboratory result case the invariant is may versus confirmed. Retain uncertainty before a follow-up analysis is completed. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because a test result must not be upgraded into proof. Retain uncertainty before a follow-up analysis is completed. If the reader believes that a technician guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 2. Product complaint: sometimes versus all the time

Source analytical gloss: The customer said the screen sometimes flickered. Unsafe rewrite: The customer admitted the screen was defective. Defensible report: The customer reported intermittent screen flicker. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a user observation, carried by a customer in communication involving a support agent. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is sometimes versus all the time. Because a service ticket can misroute diagnosis if frequency changes, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this product complaint setting, a service ticket can misroute diagnosis if frequency changes. Record intermittence and do not call it a confirmed hardware defect. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a customer and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is sometimes versus all the time. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Record intermittence and do not call it a confirmed hardware defect. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the product complaint case the invariant is sometimes versus all the time. Record intermittence and do not call it a confirmed hardware defect. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because a service ticket can misroute diagnosis if frequency changes. Record intermittence and do not call it a confirmed hardware defect. If the reader believes that a customer guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 3. Research interview: belief versus external fact

Source analytical gloss: Mina recalled, “I thought the door was locked.” Unsafe rewrite: Mina testified that the door was locked. Defensible report: Mina recalled believing the door was locked. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a remembered belief, carried by Mina in communication involving an interviewer. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is belief versus external fact. Because memory research requires separation of perception and event, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this research interview setting, memory research requires separation of perception and event. Avoid converting recollection into independent physical evidence. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to Mina and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is belief versus external fact. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Avoid converting recollection into independent physical evidence. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the research interview case the invariant is belief versus external fact. Avoid converting recollection into independent physical evidence. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because memory research requires separation of perception and event. Avoid converting recollection into independent physical evidence. If the reader believes that Mina guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 4. Public statement: claim versus verified compliance

Source analytical gloss: The spokesperson claimed the system met the standard. Unsafe rewrite: The system met the standard. Defensible report: The spokesperson asserted that the system met the standard. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a spokesperson’s assertion, carried by a spokesperson in communication involving the public. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is claim versus verified compliance. Because readers may treat organisational claims as tested results, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this public statement setting, readers may treat organisational claims as tested results. Preserve the speaker and any independent certification only if supplied. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a spokesperson and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is claim versus verified compliance. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve the speaker and any independent certification only if supplied. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the public statement case the invariant is claim versus verified compliance. Preserve the speaker and any independent certification only if supplied. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because readers may treat organisational claims as tested results. Preserve the speaker and any independent certification only if supplied. If the reader believes that a spokesperson guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 5. Novel dialogue: whispered versus declared

Source analytical gloss: “I never wanted this,” she whispered. Unsafe rewrite: She declared she was delighted. Defensible report: She quietly said she had never wanted this. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a reluctant disclosure, carried by a fictional character in communication involving a confidant. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is whispered versus declared. Because character voice and vulnerability drive the scene, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this novel dialogue setting, character voice and vulnerability drive the scene. Retain negation and the low-force manner of the utterance. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a fictional character and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is whispered versus declared. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Retain negation and the low-force manner of the utterance. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the novel dialogue case the invariant is whispered versus declared. Retain negation and the low-force manner of the utterance. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because character voice and vulnerability drive the scene. Retain negation and the low-force manner of the utterance. If the reader believes that a fictional character guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 6. Team meeting: question versus finding

Source analytical gloss: The chair asked whether the budget had been approved. Unsafe rewrite: The chair said the budget was approved. Defensible report: The chair asked if approval had been granted. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is an information-seeking question, carried by a chairperson in communication involving meeting participants. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is question versus finding. Because minutes can falsely record a decision, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this team meeting setting, minutes can falsely record a decision. Leave the approval status unresolved until a response is recorded. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a chairperson and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is question versus finding. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Leave the approval status unresolved until a response is recorded. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the team meeting case the invariant is question versus finding. Leave the approval status unresolved until a response is recorded. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because minutes can falsely record a decision. Leave the approval status unresolved until a response is recorded. If the reader believes that a chairperson guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 7. Incident log: someone versus identified actor

Source analytical gloss: A witness said, “I saw someone near the gate.” Unsafe rewrite: The witness identified the person who entered. Defensible report: The witness reported seeing an unidentified person near the gate. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a limited sighting, carried by a witness in communication involving an investigator. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is someone versus identified actor. Because evidence could be materially overstated, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this incident log setting, evidence could be materially overstated. Do not add identity or an actual entry that the account lacks. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a witness and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is someone versus identified actor. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Do not add identity or an actual entry that the account lacks. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the incident log case the invariant is someone versus identified actor. Do not add identity or an actual entry that the account lacks. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because evidence could be materially overstated. Do not add identity or an actual entry that the account lacks. If the reader believes that a witness guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 8. Medical consultation: somewhat versus completely

Source analytical gloss: The patient said the pain had eased somewhat. Unsafe rewrite: The patient confirmed the pain had stopped. Defensible report: The patient reported partial relief from pain. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a subjective symptom report, carried by a patient in communication involving a clinician. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is somewhat versus completely. Because clinical records require accurate degree of relief, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this medical consultation setting, clinical records require accurate degree of relief. Keep the patient’s experience distinct from the clinician’s assessment. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a patient and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is somewhat versus completely. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep the patient’s experience distinct from the clinician’s assessment. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the medical consultation case the invariant is somewhat versus completely. Keep the patient’s experience distinct from the clinician’s assessment. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because clinical records require accurate degree of relief. Keep the patient’s experience distinct from the clinician’s assessment. If the reader believes that a patient guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 9. Software release: expectation versus commitment

Source analytical gloss: The engineer said, “We expect the fix on Friday.” Unsafe rewrite: The engineer promised the fix would be available on Friday. Defensible report: The engineer stated that the team expected a Friday fix. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a forecast, carried by an engineer in communication involving the release manager. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is expectation versus commitment. Because planning and customer communication depend on that distinction, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this software release setting, planning and customer communication depend on that distinction. Do not turn a forecast into a guarantee. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to an engineer and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is expectation versus commitment. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Do not turn a forecast into a guarantee. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the software release case the invariant is expectation versus commitment. Do not turn a forecast into a guarantee. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because planning and customer communication depend on that distinction. Do not turn a forecast into a guarantee. If the reader believes that an engineer guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 10. Contract negotiation: conditional possibility versus accepted term

Source analytical gloss: The supplier said it could lower the price if the order increased. Unsafe rewrite: The supplier agreed to lower the price. Defensible report: The supplier said a lower price was possible if volume increased. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a conditional offer, carried by a supplier in communication involving a buyer. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is conditional possibility versus accepted term. Because legal and commercial obligations could be invented, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this contract negotiation setting, legal and commercial obligations could be invented. Keep the condition inside the report and the agreement status open. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a supplier and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is conditional possibility versus accepted term. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep the condition inside the report and the agreement status open. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the contract negotiation case the invariant is conditional possibility versus accepted term. Keep the condition inside the report and the agreement status open. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because legal and commercial obligations could be invented. Keep the condition inside the report and the agreement status open. If the reader believes that a supplier guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 11. School feedback: suggestion versus command

Source analytical gloss: The teacher suggested that the learner revise the conclusion. Unsafe rewrite: The teacher ordered the learner to rewrite the entire essay. Defensible report: The teacher recommended revising the conclusion. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is instructional advice, carried by a teacher in communication involving a learner. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is suggestion versus command. Because the nature and scope of feedback must remain accurate, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this school feedback setting, the nature and scope of feedback must remain accurate. Retain the specific writing component and advisory force. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a teacher and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is suggestion versus command. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Retain the specific writing component and advisory force. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the school feedback case the invariant is suggestion versus command. Retain the specific writing component and advisory force. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because the nature and scope of feedback must remain accurate. Retain the specific writing component and advisory force. If the reader believes that a teacher guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 12. Documentary caption: possibly versus certainty

Source analytical gloss: The archivist wrote, “Possibly taken in spring.” Unsafe rewrite: The archivist confirmed the photo was taken in spring. Defensible report: The archivist labelled the photograph as possibly taken in spring. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a tentative catalogue label, carried by an archivist in communication involving future researchers. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is possibly versus certainty. Because catalogue uncertainty must survive republication, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this documentary caption setting, catalogue uncertainty must survive republication. Preserve provenance and uncertainty markers. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to an archivist and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is possibly versus certainty. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve provenance and uncertainty markers. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the documentary caption case the invariant is possibly versus certainty. Preserve provenance and uncertainty markers. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because catalogue uncertainty must survive republication. Preserve provenance and uncertainty markers. If the reader believes that an archivist guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 13. Courtroom exchange: signing versus existence

Source analytical gloss: The witness denied signing the form. Unsafe rewrite: The witness denied the form existed. Defensible report: The witness said they had not signed the form. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a denial of signature, carried by a witness in communication involving the court. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is signing versus existence. Because denial scope shapes the disputed issue, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this courtroom exchange setting, denial scope shapes the disputed issue. Keep the exact act under negation. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a witness and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is signing versus existence. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep the exact act under negation. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the courtroom exchange case the invariant is signing versus existence. Keep the exact act under negation. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because denial scope shapes the disputed issue. Keep the exact act under negation. If the reader believes that a witness guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 14. Customer service chat: request versus warning

Source analytical gloss: The agent asked the user to restart the device. Unsafe rewrite: The agent warned the user that the device was unsafe. Defensible report: The agent requested a restart as a troubleshooting step. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a requested action, carried by a support agent in communication involving a user. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is request versus warning. Because incorrect framing can cause unnecessary alarm, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this customer service chat setting, incorrect framing can cause unnecessary alarm. Maintain purpose without inventing a danger claim. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a support agent and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is request versus warning. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Maintain purpose without inventing a danger claim. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the customer service chat case the invariant is request versus warning. Maintain purpose without inventing a danger claim. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because incorrect framing can cause unnecessary alarm. Maintain purpose without inventing a danger claim. If the reader believes that a support agent guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 15. Press conference: no evidence versus evidence of absence

Source analytical gloss: The official said, “We have no evidence of an intrusion.” Unsafe rewrite: The official proved there was no intrusion. Defensible report: The official said no evidence of intrusion had been found. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is an evidence-status report, carried by an official in communication involving reporters. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is no evidence versus evidence of absence. Because security reporting depends on this distinction, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this press conference setting, security reporting depends on this distinction. Do not infer that an intrusion was impossible. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to an official and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is no evidence versus evidence of absence. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Do not infer that an intrusion was impossible. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the press conference case the invariant is no evidence versus evidence of absence. Do not infer that an intrusion was impossible. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because security reporting depends on this distinction. Do not infer that an intrusion was impossible. If the reader believes that an official guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 16. Literary interior monologue: private cognition versus spoken challenge

Source analytical gloss: Why had he trusted the messenger? The thought returned again. Unsafe rewrite: He asked the messenger why he had trusted him. Defensible report: The narration follows the character’s private self-questioning. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a narrated thought, carried by a character in communication involving no stated listener. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is private cognition versus spoken challenge. Because a novel’s perspective would be changed, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this literary interior monologue setting, a novel’s perspective would be changed. Preserve inner questioning and avoid inventing dialogue. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a character and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is private cognition versus spoken challenge. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve inner questioning and avoid inventing dialogue. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the literary interior monologue case the invariant is private cognition versus spoken challenge. Preserve inner questioning and avoid inventing dialogue. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because a novel’s perspective would be changed. Preserve inner questioning and avoid inventing dialogue. If the reader believes that a character guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 17. Policy review: recommendation versus adoption

Source analytical gloss: The panel recommended further assessment before adoption. Unsafe rewrite: The panel approved immediate adoption. Defensible report: The panel recommended a further assessment before any adoption decision. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a procedural recommendation, carried by a review panel in communication involving a governing body. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is recommendation versus adoption. Because governance records must separate advice from decisions, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this policy review setting, governance records must separate advice from decisions. Preserve the sequence and lack of approval. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a review panel and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is recommendation versus adoption. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve the sequence and lack of approval. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the policy review case the invariant is recommendation versus adoption. Preserve the sequence and lack of approval. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because governance records must separate advice from decisions. Preserve the sequence and lack of approval. If the reader believes that a review panel guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 18. Research paper: might reflect versus established cause

Source analytical gloss: The authors argued that the pattern might reflect selection bias. Unsafe rewrite: The authors established that selection bias caused the pattern. Defensible report: The authors proposed selection bias as a possible explanation. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a qualified interpretation, carried by research authors in communication involving scholarly readers. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is might reflect versus established cause. Because scientific uncertainty should not be erased, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this research paper setting, scientific uncertainty should not be erased. Retain modal language and distinguish explanation from proof. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to research authors and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is might reflect versus established cause. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Retain modal language and distinguish explanation from proof. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the research paper case the invariant is might reflect versus established cause. Retain modal language and distinguish explanation from proof. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because scientific uncertainty should not be erased. Retain modal language and distinguish explanation from proof. If the reader believes that research authors guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 19. Oral history: grandmother’s report versus participant’s observation

Source analytical gloss: The participant said, “My grandmother told me the bridge once collapsed.” Unsafe rewrite: The participant witnessed the bridge collapse. Defensible report: The participant passed on a family account of an earlier collapse. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a nested oral account, carried by a participant in communication involving an oral historian. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is grandmother’s report versus participant’s observation. Because provenance may be lost across generations, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this oral history setting, provenance may be lost across generations. Keep both the intermediary and the original claim. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a participant and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is grandmother’s report versus participant’s observation. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep both the intermediary and the original claim. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the oral history case the invariant is grandmother’s report versus participant’s observation. Keep both the intermediary and the original claim. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because provenance may be lost across generations. Keep both the intermediary and the original claim. If the reader believes that a participant guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 20. Film subtitle: maybe versus certainty

Source analytical gloss: He muttered, “Maybe they forgot.” Unsafe rewrite: He declared, “They forgot.” Defensible report: He quietly speculated that they might have forgotten. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is an uncertain line, carried by a film character in communication involving another character. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is maybe versus certainty. Because subtitles may compress but must retain uncertainty, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this film subtitle setting, subtitles may compress but must retain uncertainty. Use a short target phrase preserving speculation and manner. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a film character and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is maybe versus certainty. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Use a short target phrase preserving speculation and manner. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the film subtitle case the invariant is maybe versus certainty. Use a short target phrase preserving speculation and manner. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because subtitles may compress but must retain uncertainty. Use a short target phrase preserving speculation and manner. If the reader believes that a film character guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 21. Workplace feedback: some versus all

Source analytical gloss: The manager said, “Some deadlines were missed.” Unsafe rewrite: The manager accused everyone of missing all deadlines. Defensible report: The manager said that some deadlines had not been met. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a limited performance observation, carried by a manager in communication involving a team. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is some versus all. Because an overgeneralised report can damage trust, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this workplace feedback setting, an overgeneralised report can damage trust. Keep the quantity and do not invent individual culpability. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a manager and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is some versus all. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep the quantity and do not invent individual culpability. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the workplace feedback case the invariant is some versus all. Keep the quantity and do not invent individual culpability. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because an overgeneralised report can damage trust. Keep the quantity and do not invent individual culpability. If the reader believes that a manager guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 22. Safety briefing: future risk versus present status

Source analytical gloss: The officer warned that the route could become unsafe in heavy rain. Unsafe rewrite: The officer announced the route was already unsafe. Defensible report: The officer warned of possible danger during heavy rain. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a conditional safety warning, carried by a safety officer in communication involving visitors. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is future risk versus present status. Because poor translation could alter route choices, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this safety briefing setting, poor translation could alter route choices. Retain the weather condition and potential, not current certainty. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a safety officer and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is future risk versus present status. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Retain the weather condition and potential, not current certainty. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the safety briefing case the invariant is future risk versus present status. Retain the weather condition and potential, not current certainty. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because poor translation could alter route choices. Retain the weather condition and potential, not current certainty. If the reader believes that a safety officer guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 23. Memoir: believed versus fact

Source analytical gloss: She wrote, “I believed my brother had already left.” Unsafe rewrite: She documented that her brother had already left. Defensible report: She wrote that she had believed her brother was already gone. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a retrospective belief, carried by a memoirist in communication involving readers. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is believed versus fact. Because first-person memory and external reality must remain distinct, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this memoir setting, first-person memory and external reality must remain distinct. Keep the narrator’s past epistemic state. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a memoirist and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is believed versus fact. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep the narrator’s past epistemic state. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the memoir case the invariant is believed versus fact. Keep the narrator’s past epistemic state. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because first-person memory and external reality must remain distinct. Keep the narrator’s past epistemic state. If the reader believes that a memoirist guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 24. Broadcast correction: initial claim versus repaired claim

Source analytical gloss: The host first said the vote was unanimous, then corrected it to a majority. Unsafe rewrite: The host maintained the vote was unanimous. Defensible report: The host corrected an earlier unanimity claim to a majority result. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a self-correction, carried by a broadcaster in communication involving viewers. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is initial claim versus repaired claim. Because public records should show correction history when material, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this broadcast correction setting, public records should show correction history when material. Preserve the repair rather than erasing it. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a broadcaster and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is initial claim versus repaired claim. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve the repair rather than erasing it. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the broadcast correction case the invariant is initial claim versus repaired claim. Preserve the repair rather than erasing it. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because public records should show correction history when material. Preserve the repair rather than erasing it. If the reader believes that a broadcaster guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 25. Survey response: consider versus commit

Source analytical gloss: The respondent wrote, “I would consider subscribing.” Unsafe rewrite: The respondent promised to subscribe. Defensible report: The respondent expressed conditional interest in subscribing. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a hypothetical intention, carried by a respondent in communication involving a researcher. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is consider versus commit. Because survey interpretation must not inflate conversion intent, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this survey response setting, survey interpretation must not inflate conversion intent. Retain conditionality and the absence of a definite purchase. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a respondent and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is consider versus commit. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Retain conditionality and the absence of a definite purchase. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the survey response case the invariant is consider versus commit. Retain conditionality and the absence of a definite purchase. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because survey interpretation must not inflate conversion intent. Retain conditionality and the absence of a definite purchase. If the reader believes that a respondent guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 26. Academic citation: expert voice versus article conclusion

Source analytical gloss: The article quoted an expert who called the proposal ‘premature’. Unsafe rewrite: The article itself proved the proposal premature. Defensible report: The article attributed the word ‘premature’ to a quoted expert. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is an embedded quotation, carried by an expert cited by an article in communication involving the article’s readers. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is expert voice versus article conclusion. Because source layering matters in scholarship, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this academic citation setting, source layering matters in scholarship. Keep the fragment with its original speaker. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to an expert cited by an article and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is expert voice versus article conclusion. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep the fragment with its original speaker. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the academic citation case the invariant is expert voice versus article conclusion. Keep the fragment with its original speaker. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because source layering matters in scholarship. Keep the fragment with its original speaker. If the reader believes that an expert cited by an article guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 27. Community meeting: request for date versus declaration of date

Source analytical gloss: A resident asked when repairs would begin. Unsafe rewrite: A resident announced repairs would begin immediately. Defensible report: A resident requested information about the repair schedule. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a timing question, carried by a resident in communication involving meeting officials. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is request for date versus declaration of date. Because minutes should not invent operational commitments, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this community meeting setting, minutes should not invent operational commitments. Preserve the unresolved question and exact project referent. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a resident and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is request for date versus declaration of date. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve the unresolved question and exact project referent. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the community meeting case the invariant is request for date versus declaration of date. Preserve the unresolved question and exact project referent. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because minutes should not invent operational commitments. Preserve the unresolved question and exact project referent. If the reader believes that a resident guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 28. Financial disclosure: forecast versus historical fact

Source analytical gloss: The report states that management expects lower costs next year. Unsafe rewrite: The report verifies that costs will fall next year. Defensible report: The report attributes a forecast of lower costs to management. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is an attributed forecast, carried by management in communication involving investors. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is forecast versus historical fact. Because financial readers rely on forward-looking distinctions, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this financial disclosure setting, financial readers rely on forward-looking distinctions. Preserve the authorising source and uncertainty. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to management and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is forecast versus historical fact. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve the authorising source and uncertainty. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the financial disclosure case the invariant is forecast versus historical fact. Preserve the authorising source and uncertainty. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because financial readers rely on forward-looking distinctions. Preserve the authorising source and uncertainty. If the reader believes that management guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 29. Disciplinary hearing: receipt versus consent

Source analytical gloss: The employee acknowledged receiving the message but denied agreeing to the instruction. Unsafe rewrite: The employee admitted agreeing to the instruction. Defensible report: The employee accepted receipt while denying agreement. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a two-part stance, carried by an employee in communication involving hearing participants. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is receipt versus consent. Because partial acknowledgement must not become admission of obligation, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this disciplinary hearing setting, partial acknowledgement must not become admission of obligation. Translate each verb and its object separately. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to an employee and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is receipt versus consent. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Translate each verb and its object separately. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the disciplinary hearing case the invariant is receipt versus consent. Translate each verb and its object separately. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because partial acknowledgement must not become admission of obligation. Translate each verb and its object separately. If the reader believes that an employee guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 30. Museum label: attributed authorship versus verified authorship

Source analytical gloss: A letter attributed to the explorer says the journey began at dawn. Unsafe rewrite: The explorer definitely wrote that the journey began at dawn. Defensible report: A letter attributed to the explorer reports a dawn departure. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is an uncertain document attribution, carried by an archival source in communication involving museum visitors. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is attributed authorship versus verified authorship. Because historical collections require careful provenance, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this museum label setting, historical collections require careful provenance. Preserve the document’s disputed authorship status. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to an archival source and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is attributed authorship versus verified authorship. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve the document’s disputed authorship status. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the museum label case the invariant is attributed authorship versus verified authorship. Preserve the document’s disputed authorship status. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because historical collections require careful provenance. Preserve the document’s disputed authorship status. If the reader believes that an archival source guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 31. Parent communication: possibility versus actual cancellation

Source analytical gloss: The coordinator wrote that the trip might be postponed if conditions deteriorated. Unsafe rewrite: The coordinator cancelled the trip. Defensible report: The coordinator advised that postponement remained possible under worsening conditions. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a contingency notice, carried by a coordinator in communication involving families. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is possibility versus actual cancellation. Because unnecessary travel or anxiety can result from altered action status, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this parent communication setting, unnecessary travel or anxiety can result from altered action status. Keep the condition and pending decision distinct. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a coordinator and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is possibility versus actual cancellation. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep the condition and pending decision distinct. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the parent communication case the invariant is possibility versus actual cancellation. Keep the condition and pending decision distinct. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because unnecessary travel or anxiety can result from altered action status. Keep the condition and pending decision distinct. If the reader believes that a coordinator guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 32. Technical manual: after versus caused by

Source analytical gloss: The reviewer noted that the sensor appeared to drift after calibration. Unsafe rewrite: The reviewer confirmed that calibration caused sensor failure. Defensible report: The reviewer observed possible sensor drift following calibration. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a tentative observation, carried by a reviewer in communication involving engineers. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is after versus caused by. Because technical troubleshooting depends on causal restraint, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this technical manual setting, technical troubleshooting depends on causal restraint. Preserve order, appearance and non-established causation. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a reviewer and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is after versus caused by. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve order, appearance and non-established causation. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the technical manual case the invariant is after versus caused by. Preserve order, appearance and non-established causation. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because technical troubleshooting depends on causal restraint. Preserve order, appearance and non-established causation. If the reader believes that a reviewer guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 33. Humanitarian interview: reported community information versus direct measurement

Source analytical gloss: The interviewee said neighbours had told her that water was scarce. Unsafe rewrite: The interviewee measured low water supplies herself. Defensible report: The interviewee relayed neighbours’ accounts of water scarcity. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a secondary report, carried by an interviewee in communication involving field researchers. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is reported community information versus direct measurement. Because aid planning needs transparent evidence, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this humanitarian interview setting, aid planning needs transparent evidence. Keep the evidential chain and avoid inventing a measurement. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to an interviewee and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is reported community information versus direct measurement. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep the evidential chain and avoid inventing a measurement. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the humanitarian interview case the invariant is reported community information versus direct measurement. Keep the evidential chain and avoid inventing a measurement. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because aid planning needs transparent evidence. Keep the evidential chain and avoid inventing a measurement. If the reader believes that an interviewee guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 34. Fictional narrator: free indirect optimism versus narrator fact

Source analytical gloss: Of course he was innocent. The council would see that tomorrow, he thought. Unsafe rewrite: The narrator established his innocence and tomorrow’s verdict. Defensible report: The character believed the council would vindicate him the next day. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a character-centred expectation, carried by a fictional character in communication involving the reader. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is free indirect optimism versus narrator fact. Because narrative irony may depend on the gap, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this fictional narrator setting, narrative irony may depend on the gap. Do not collapse a viewpoint into objective narrative certainty. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a fictional character and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is free indirect optimism versus narrator fact. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Do not collapse a viewpoint into objective narrative certainty. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the fictional narrator case the invariant is free indirect optimism versus narrator fact. Do not collapse a viewpoint into objective narrative certainty. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because narrative irony may depend on the gap. Do not collapse a viewpoint into objective narrative certainty. If the reader believes that a fictional character guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 35. Interpreter handover: requested clarification versus disclaimed replacement

Source analytical gloss: The caller said, “Please tell the specialist I need clarification, not a replacement.” Unsafe rewrite: The caller ordered a replacement. Defensible report: The caller asked the specialist for clarification rather than a replacement. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a mediated request, carried by a caller in communication involving a specialist via an interpreter. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is requested clarification versus disclaimed replacement. Because service outcomes depend on the exact object of the request, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this interpreter handover setting, service outcomes depend on the exact object of the request. Keep the contrast and both roles visible. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a caller and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is requested clarification versus disclaimed replacement. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep the contrast and both roles visible. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the interpreter handover case the invariant is requested clarification versus disclaimed replacement. Keep the contrast and both roles visible. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because service outcomes depend on the exact object of the request. Keep the contrast and both roles visible. If the reader believes that a caller guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 36. Legal interview: reportedly versus first-hand certainty

Source analytical gloss: The interview note says, ‘He reportedly saw a blue vehicle.’ Unsafe rewrite: He saw a blue vehicle and identified it. Defensible report: The note gives a second-hand account that a blue vehicle may have been seen. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a reported visual account, carried by an unnamed observer in communication involving an investigator. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is reportedly versus first-hand certainty. Because evidence chain and identity should not be invented, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this legal interview setting, evidence chain and identity should not be invented. Maintain second-hand status and limited description. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to an unnamed observer and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is reportedly versus first-hand certainty. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Maintain second-hand status and limited description. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the legal interview case the invariant is reportedly versus first-hand certainty. Maintain second-hand status and limited description. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because evidence chain and identity should not be invented. Maintain second-hand status and limited description. If the reader believes that an unnamed observer guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 37. Peer review: question versus definitive criticism

Source analytical gloss: The reviewer asked whether the sample size was sufficient. Unsafe rewrite: The reviewer said the sample size was insufficient. Defensible report: The reviewer questioned the adequacy of the sample size. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a methodological question, carried by a reviewer in communication involving authors. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is question versus definitive criticism. Because research feedback may be exploratory rather than conclusive, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this peer review setting, research feedback may be exploratory rather than conclusive. Preserve the question’s openness. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a reviewer and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is question versus definitive criticism. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve the question’s openness. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the peer review case the invariant is question versus definitive criticism. Preserve the question’s openness. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because research feedback may be exploratory rather than conclusive. Preserve the question’s openness. If the reader believes that a reviewer guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 38. Apology dispute: regret versus liability admission

Source analytical gloss: The representative said, “We regret the delay.” Unsafe rewrite: The representative admitted legal responsibility for the delay. Defensible report: The representative expressed regret about the delay. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is an expression of regret, carried by a representative in communication involving customers. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is regret versus liability admission. Because organisational language can have distinct legal and relational effects, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this apology dispute setting, organisational language can have distinct legal and relational effects. Do not insert an admission absent from the original. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a representative and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is regret versus liability admission. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Do not insert an admission absent from the original. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the apology dispute case the invariant is regret versus liability admission. Do not insert an admission absent from the original. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because organisational language can have distinct legal and relational effects. Do not insert an admission absent from the original. If the reader believes that a representative guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 39. Retraction notice: withdrawal versus persistence

Source analytical gloss: The researcher withdrew an earlier claim of statistical significance. Unsafe rewrite: The researcher repeated the significance claim. Defensible report: The researcher retracted the earlier significance claim. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is a formal retraction, carried by a researcher in communication involving research community. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is withdrawal versus persistence. Because knowledge correction must remain visible, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this retraction notice setting, knowledge correction must remain visible. Preserve reversal and record the earlier position accurately. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a researcher and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is withdrawal versus persistence. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Preserve reversal and record the earlier position accurately. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the retraction notice case the invariant is withdrawal versus persistence. Preserve reversal and record the earlier position accurately. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because knowledge correction must remain visible. Preserve reversal and record the earlier position accurately. If the reader believes that a researcher guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

Case 40. Negotiation call: possibility query versus ultimatum

Source analytical gloss: The buyer asked, ‘Could delivery happen sooner?’ Unsafe rewrite: The buyer demanded expedited delivery. Defensible report: The buyer asked whether earlier delivery was possible. The three formulations are not stylistic alternatives. The first presents a source-language reporting arrangement; the second changes the accountable event or commitment; the third is a proposed meaning-preserving reconstruction that still needs natural expression in the actual target language.

Event reconstruction. The core event is an exploratory request, carried by a buyer in communication involving a supplier. A translator begins by recording that scene rather than translating the most eye-catching reporting verb. The essential distinction is possibility query versus ultimatum. Because commercial relationships change with invented pressure, a seemingly minor change of expression can redirect the reader’s judgement about whether something was asked, observed, inferred, denied, promised or verified. Before drafting, write the minimum content that survives the event and the minimum source frame that must accompany it. That record protects the speaker even if target grammar reorders the sentence. Diagnostic task: Write a neutral one-sentence event record that mentions the source, communicative act and its limits; check that it does not introduce information absent from the original.

Boundary comparison. Compare the source and unsafe rewrite word by word only after locating their reporting frames. The unsafe form moves attention toward a different responsibility relation, while the defensible version keeps the original boundary intact. In this negotiation call setting, commercial relationships change with invented pressure. Keep the request tentative and without an imposed deadline. The translator must examine whether quotation marks are signalling exact words, whether an indirect clause is summarising content, and whether a descriptive editorial phrase could be mistaken for the source speaker’s own language. The output must tell the target reader which claim belongs to a buyer and which explanation belongs to the later writer. Diagnostic task: Circle the reporting expression in each version. Then label the proposition it governs and identify any newly asserted proposition.

Verb force and commitment. Test the reporting verb as though it were a claim about the event. The verbs in the source and unsafe rewrite are not interchangeable when the active contrast is possibility query versus ultimatum. If a target-language verb has broader meaning than the English analytical gloss, inspect the surrounding clause, particles, aspect, honorifics or evidential markings before deciding. A correct translation may need more than one word to preserve one reporting act. Keep the request tentative and without an imposed deadline. The question is not which verb sounds strongest; it is which verb can be defended by the source record without importing motive, certainty, authority or agreement. Diagnostic task: Replace the target reporting verb with a neutral ‘said’ or ‘reported’; determine exactly what speech-act information is lost and what must be restored.

Perspective, deixis and sequence. Now move the reporting event to a different moment and viewpoint without changing its content. The original words may contain person references or implicit time anchors that the later report must recalculate; conversely, an actual direct quote should normally maintain the speaker’s original viewpoint unless the translation brief licenses a marked intervention. In the negotiation call case the invariant is possibility query versus ultimatum. Keep the request tentative and without an imposed deadline. Identify what was true or believed at the original time and what the reporter knows later. Do not let grammatical backshift manufacture a completed event, and do not automatically keep a present tense if it falsely suggests an ongoing state. Diagnostic task: Imagine the report is read one year later. Rewrite only the necessary time or person references and explain why the speaker’s original commitment is unchanged.

Independent-reader diagnostic. Hide the source and present the safe target to a reader who has no access to the surrounding explanation. Ask who said the content, how the statement reached the record, and whether it is confirmed, questioned, denied, predicted, remembered or merely attributed. Then show the source and compare the answers. This is crucial because commercial relationships change with invented pressure. Keep the request tentative and without an imposed deadline. If the reader believes that a buyer guaranteed or personally witnessed a fact that was only reported provisionally, reopen the case. A fluent sentence is not an accurate translation when its responsibility pattern invites an unsupported answer. Diagnostic task: Ask a second reviewer to write two separate sentences: ‘The source reports that …’ and ‘The document itself asserts that …’. Compare whether both can be justified.

7. Thirty recurring failure modes and targeted repairs

Failure 1: Quotation inflation

How it appears: Paraphrased content gains quotation marks. Why it matters: Readers mistake reconstructed phrasing for directly attested words. Repair: Keep quotation scope tied to evidence, or identify the passage as a summary. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 2: Attribution evaporation

How it appears: A named source disappears during sentence compression. Why it matters: A source’s contested view becomes narrator fact. Repair: Restore the reporting frame near the claim. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 3: Unjustified confession

How it appears: Said becomes admitted without a source concession. Why it matters: The target invents resistance, guilt or acceptance. Repair: Choose a neutral speech verb until confession is explicit. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 4: False verification

How it appears: Claimed becomes confirmed or demonstrated. Why it matters: A disputed assertion looks verified. Repair: Separate source assertion from evidence. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 5: Misplaced addressee

How it appears: A target reporting construction invents the recipient. Why it matters: The sentence changes the social event. Repair: Check who was actually addressed. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 6: Nested-speaker collapse

How it appears: A reports B quoting C but C disappears. Why it matters: Readers assign testimony to the wrong participant. Repair: Number frames and preserve intermediary markers. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 7: Thought made audible

How it appears: Internal wondering becomes a spoken question. Why it matters: A fictional or biographical event is invented. Repair: Use thought or perception language where appropriate. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 8: Question asserted

How it appears: A reported question is rewritten as its positive answer. Why it matters: Unresolved information becomes a claim. Repair: Retain interrogative force in the report. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 9: Request upgraded

How it appears: A suggestion becomes a command. Why it matters: Authority is exaggerated. Repair: Separate advisory force from obligatory force. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 10: Warning turned threat

How it appears: Warning of danger becomes threatening to cause danger. Why it matters: The speaker’s agency reverses. Repair: Ask who could cause the harm. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 11: Forecast made promise

How it appears: Expected becomes guaranteed. Why it matters: The source appears to undertake a commitment. Repair: Restore uncertainty and conditional terms. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 12: Denial scope expansion

How it appears: A person denies one action but seems to deny the whole incident. Why it matters: The record changes what is contested. Repair: Underline the exact predicate under negation. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 13: Partial concession flattened

How it appears: The speaker accepts a detail but not a conclusion. Why it matters: A limited acknowledgement becomes total agreement. Repair: Retain both conceded and resisted propositions. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 14: Backshift by reflex

How it appears: A tense is altered mechanically. Why it matters: Event chronology becomes wrong. Repair: Reconstruct three times: speech, report and event. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 15: Time deictic drift

How it appears: Tomorrow is copied after a reporting-date shift. Why it matters: A deadline moves silently. Repair: Anchor relative expressions to a known date. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 16: Pronoun ownership drift

How it appears: I or we changes referent in a nested report. Why it matters: Credit and responsibility move to someone else. Repair: Build a speaker-index map. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 17: Evidential deletion

How it appears: Apparently or reportedly disappears. Why it matters: Hearsay becomes direct observation. Repair: Preserve the evidence channel. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 18: Overconfident translation

How it appears: A hesitant statement becomes unequivocal. Why it matters: The target increases claim certainty. Repair: Match the original commitment level. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 19: Invented listener

How it appears: A speech act acquires a named recipient. Why it matters: A social interaction is fabricated. Repair: Keep missing participants unspecified. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 20: Quotation punctuation imitation

How it appears: Source punctuation is copied despite target convention. Why it matters: Reader misparses speech boundaries. Repair: Use target conventions while maintaining exact scope. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 21: Reporter endorsement

How it appears: A citation frame becomes endorsement of its proposition. Why it matters: Source trust is overstated. Repair: Distinguish ‘the article quotes’ from ‘the article concludes’. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 22: Headline strengthening

How it appears: A compressed headline removes ‘according to’. Why it matters: Readers treat an allegation as established fact. Repair: Keep the minimum attribution marker. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 23: Fragment laundering

How it appears: A partial quote is made into a full sentence. Why it matters: A quote may acquire a new intended meaning. Repair: Signal grammatical adaptation or paraphrase. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 24: Source-language gloss mistaken for target form

How it appears: A bilingual working gloss is published literally. Why it matters: The target sounds unnatural or carries wrong syntax. Repair: Rebuild idiomatic grammar after preserving meaning. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 25: Speaker-switch error

How it appears: A dialogue line is attached to a neighbouring speaker. Why it matters: Character identities or responsibilities reverse. Repair: Track each speech turn across paragraph transitions. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 26: Missing retraction

How it appears: The corrected version replaces all traces of an earlier claim. Why it matters: A documented repair disappears. Repair: Preserve correction when relevant to the record. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 27: Genre mismatch

How it appears: An oral history is edited like a legal transcript or vice versa. Why it matters: Evidence and voice conventions are violated. Repair: Apply the commission’s quotation policy. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 28: Consent inferred

How it appears: Receiving a message becomes agreeing to its terms. Why it matters: An obligation is invented. Repair: Distinguish acknowledgement, understanding and acceptance. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 29: Silence read as assent

How it appears: No answer is summarised as agreement. Why it matters: A non-event becomes affirmative evidence. Repair: Report the silence or non-response. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

Failure 30: Indirect report over-explicated

How it appears: The translator explains a possible motive as though sourced. Why it matters: Readers cannot tell text from interpretation. Repair: Keep commentary visibly outside the reported claim. The editor should point to an identifiable source span and an identifiable target span, then test which reader inference changed. If the difference is only a legitimate target-language grammatical form, leave it alone. If the difference moves an utterance to a new speaker, changes the evidence status, expands a denial or changes the communicative force, classify it as a meaning error and correct it. Document the fix so an automatic rewriting pass cannot reintroduce the stronger reading.

8. Twelve transfer laboratories for translators, editors and language learners

Laboratory 1: The interview transcript

Select a recorded interview with three speakers. Mark exact quoted spans, uncertain words, interruptions and paraphrases separately. Write a translation where the interviewer’s summarising language remains distinguishable from the participant’s actual responses. Evaluate whether a reader could recover each speaker’s position without consulting the audio; when recording quality is poor, preserve uncertainty rather than fabricating completeness. Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

Laboratory 2: The multilingual newsroom

Take a short article that reports an official claim and an independent criticism. Translate both without making the first source the narrator’s truth or the second source a proven rebuttal. Check headline, opening paragraph, captions and pull quotes for lost attribution. Different page positions can require their own attribution cues. Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

Laboratory 3: The fictional conversation

Translate six alternating dialogue turns that include a whisper, an unspoken thought and an interrupted answer. Preserve which lines are audible. Ask a second reader to identify who actually knows what, because narrator omniscience and character speculation must not merge. Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

Laboratory 4: The customer complaint

Convert customer messages into a service ticket while retaining frequency, degree and source of each observation. Avoid translating ‘sometimes’ into ‘always’ or ‘I think’ into ‘I know’. Test whether the technician can distinguish confirmed defects from the customer’s descriptions and planned diagnostics. Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

Laboratory 5: The institutional minutes

Translate minutes containing questions, proposals, decisions, and unresolved points. Use headings only if the source document’s format permits. Audit whether every ‘agreed’, ‘asked’ and ‘recommended’ entry has the correct speaker, audience and action status. Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

Laboratory 6: The medical narrative

Translate an anonymised account in which a patient reports pain, a caregiver supplies additional context, and a clinician records an examination. Create separate source tags for symptom report, collateral account and clinical finding; do not confuse one with another. Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

Laboratory 7: The academic abstract

Translate a paragraph with author hypotheses, prior-work quotations and reported findings. Retain hedges and references. Check that phrases such as ‘suggests’, ‘supports’ and ‘demonstrates’ have not converged into a single overly certain verb. Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

Laboratory 8: The historical exhibit

Translate a museum panel quoting a letter whose authorship is tentative. Preserve attribution uncertainty and document title conventions. Separate the curator’s assessment from what the historical source actually says. Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

Laboratory 9: The safety instruction

Translate a warning that quotes a manufacturer’s limitation and an operator’s local procedure. Preserve which party owns each directive. Run an action test: does the reader know what must be done now and what is only cited background? Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

Laboratory 10: The subtitled documentary

Translate a narrated claim with a direct interview extract under subtitle constraints. Shorten responsibly without moving the quote into the narrator’s mouth. Test the subtitles against the audio to see whether timing changes perceived attribution. Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

Laboratory 11: The family recollection

Translate an account where one person recalls what a relative said decades earlier. Keep remembered speech distinct from verified events. Check whether the target expresses layered memory and whether any invented precision makes the recollection misleading. Complete the practice in three versions: a structurally close draft, a natural target-language draft and a final verified version. After each version, answer four questions: who is responsible for the reported content; what evidence status does the text convey; what communicative act occurred; and what remains uncertain? A good exercise produces not only an acceptable sentence but an explicit explanation of why an alternative changes the source event. Keep a log of the error you nearly made; the resulting examples become a personal translation-quality reference.

9. A four-week progression from recognising reports to independently defending them

Week 1: recover the event

Work on short sentences and label speaker, addressee, proposition, reporting verb and quote span. Every day choose five contrasting reports and explain the difference without using a bilingual dictionary first. At the end of the week, ask whether a reader can distinguish quotation, indirect report and narrator interpretation in twenty fresh examples. Do not grade only the number of fluent sentences. Grade preservation of attribution, quotation status, force, epistemic caution and reader action. A mismatch in a high-consequence passage merits more attention than a minor difference in punctuation. Record before-and-after examples so the work accumulates into a reusable method, rather than becoming another set of corrected exercises that disappear after review.

Week 2: reconstruct perspective

Practise pronoun referents, time anchors, nested speakers and questions. Move the same utterance between direct and indirect reporting where the task licenses it, then reverse the operation. Measure success by whether all original participants, conditions and commitments remain identifiable after each transformation. Do not grade only the number of fluent sentences. Grade preservation of attribution, quotation status, force, epistemic caution and reader action. A mismatch in a high-consequence passage merits more attention than a minor difference in punctuation. Record before-and-after examples so the work accumulates into a reusable method, rather than becoming another set of corrected exercises that disappear after review.

Week 3: work across genres

Translate small sections of fictional dialogue, journalism, interviews, minutes and informational writing. Use the relevant genre’s quotation conventions without altering source responsibility. For every genre, record one permissible adaptation and one impermissible strengthening, with a justification tied to the source. Do not grade only the number of fluent sentences. Grade preservation of attribution, quotation status, force, epistemic caution and reader action. A mismatch in a high-consequence passage merits more attention than a minor difference in punctuation. Record before-and-after examples so the work accumulates into a reusable method, rather than becoming another set of corrected exercises that disappear after review.

Week 4: independent verification

Choose a complete short text, draft it, perform a reporting-boundary review and give the target alone to a reader. Record what that reader thinks the speaker said. Compare the reader’s interpretation with the source and refine the version. A clean independent read is stronger evidence of quality than familiarity with your own translation. Do not grade only the number of fluent sentences. Grade preservation of attribution, quotation status, force, epistemic caution and reader action. A mismatch in a high-consequence passage merits more attention than a minor difference in punctuation. Record before-and-after examples so the work accumulates into a reusable method, rather than becoming another set of corrected exercises that disappear after review.

10. Frequently asked questions about translating reported speech

What is the difference between direct and indirect speech in translation?

Direct reporting presents a speaker’s words as quoted wording, while indirect reporting carries a speech event’s content or force without the same exact-word status. Translation must preserve that distinction even if quotation punctuation or grammar changes across languages. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

Does every indirect report require English backshift?

No. English allows several patterns depending on time, continuing relevance and reporting situation; other languages organise speech time differently. Establish the actual event chronology and target convention instead of applying a chart mechanically. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

Can said and told be translated by one verb?

Sometimes a target language uses one general reporting verb with different constructions. What matters is whether the original identifies an addressee and whether the target construction preserves that social and grammatical relation. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

Why is claimed different from proved?

Claimed locates an assertion with a source; proved represents a much stronger epistemic result. Conflating them can turn contested testimony into what appears to be independently established fact. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

Should translated direct quotes stay in quotation marks?

Follow the commission and target-language genre conventions. Keep the span of words attributed to the speaker clear and avoid using marks to suggest verbatim fidelity where the source only provides a paraphrase. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

What about ‘according to’ statements?

According to introduces a source frame. If deleted, the target may appear to endorse the source’s claim. A target-language equivalent can be syntactically different but should make the attribution recoverable. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

How are reported questions translated?

Identify whether the question seeks information, performs a request or functions rhetorically. Rebuild it with target-language indirect-question grammar while keeping its interrogative force and referents. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

What changes when quoting an earlier translation?

A translation of speech is already an interpretation of wording in another language. Editorial notes may need to distinguish a source-language quotation, an authorised translation and a later paraphrase. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

Can a translator simplify complex reporting verbs?

Yes when the target language requires a different expression, but the lost information must be restored where material. Replacing confessed with said can erase admission; replacing said with confessed invents it. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

How does free indirect discourse affect translation?

It can blend narrator syntax with a character’s thoughts or evaluations. Translating it as plain narrated fact or as directly spoken dialogue can destroy narrative perspective and irony. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

What if the original speaker is unknown?

Keep the uncertainty visible. Use an appropriately unspecified source or report marker rather than inventing a named witness, authority or participant. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

Can a report have several speakers inside one sentence?

Yes. Nested quotes and hearsay chains are common. Create a speaker map and mark each layer before drafting. The number of reporting layers is part of the information. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

How should denials be translated?

Record the precise proposition that is denied. ‘I did not sign’ differs from ‘I did not receive’ and ‘I did not agree’; a fluent but broader denial changes the record. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

Can reporting verbs change tone?

Yes. Complained, insisted, snapped, whispered, argued and admitted carry different implications. Use them when supported, not as decorative synonyms intended to avoid repetition. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

How do I handle unfinished quotations?

Respect the genre’s rules for interruptions, ellipses and editorial intervention. If a source utterance is incomplete, avoid making its eventual conclusion sound certain. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

Is reported speech translation only an English grammar topic?

No. Every language has resources for referring to others’ utterances, but structures and evidential systems differ. The general problem is preserving source responsibility; English grammar is only one possible implementation. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

How do I check a machine-translated report?

Make a ledger of all reporting verbs, quotes and speaker names. Check changes of certainty, direction of denial, nested source chains and dates. Fluency alone is insufficient. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

Where do vocabulary lessons fit?

Vocabulary learning helps distinguish speech-act verbs and their collocations. The protected Vocabulary Learning Hub owns learning and retrieval; this article owns how those distinctions survive a cross-language report. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

Where do English grammar explanations fit?

How English Works explains the internal resources of English. This article applies those resources in translation while also asking how non-English languages represent the same communication event. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

What is the fastest practical check before release?

Hide the source and ask a target reader: who asserted what, how was it known, was it said aloud, and what remained uncertain? Reopen any sentence whose answer differs from the source record. For important work, make the answer explicit in the translation note: identify the original span, the target choice and the reader inference you intend to preserve. If alternative phrasings differ in whether they assert, question, doubt, deny, promise or merely report a proposition, that is a meaning decision rather than a surface preference.

11. Quality assurance: classify real errors rather than simply marking sentences as awkward

A useful revision pass separates accuracy, terminology, language conventions, style and audience appropriateness. The MQM Council’s translation-quality typology describes accuracy problems such as mistranslation, omission and addition. Applied here, an omitted reporting verb can become an attribution error; an added commitment word can become an overstatement; a shifted pronoun can become a role error. These categories help teams talk about consequences rather than calling every concern ‘bad English’. They are diagnostic tools, not substitutes for subject expertise or professional review where a text has legal, medical, safety or other high-stakes implications.

Report the defect with evidence. For example: source ‘the engineer expects completion’; target ‘the engineer guarantees completion’; issue ‘an uncertain forecast became a firm promise’; risk ‘readers may rely on a false undertaking’; correction ‘restore an expectation marker’. This format travels across language pairs and can be applied by human editors to machine-translated drafts. If two reviewers disagree, have them separately explain the communicative act in source and target before debating word preferences.

12. What this chapter owns — and what its neighbouring owners do not need to duplicate

This article owns cross-language reconstruction of reported speech boundaries: what was directly quoted, what was reported indirectly, who carried the speech act, what responsibility the report attributes, and which changes in target grammar preserve that relation. For the larger method of moving meaning across languages, return to Master Art of Translation | The Complete System for Moving Meaning Between Languages. The previous specialised chapter, the Causal Attribution Fidelity System, focuses on preserving reasons and results; this chapter instead asks who is responsible for an utterance and what kind of utterance occurred. A causal claim can itself be quoted or reported, but the two diagnostic jobs should remain distinct.

When the difficulty is acquiring or choosing words such as concede, insist, deny, claim or report in ordinary reading and writing, use the protected Vocabulary Learning Hub rather than reproducing its teaching system here. When the difficulty is understanding the grammar of English indirect questions, clause structure or register inside English, use the existing How English Works owner. These outward routes are relevant only to their own jobs; the present article remains a world-facing guide to translation across any language pair.

13. Research references and further reading

• Cambridge Dictionary, Reported speech and reporting verbs — distinguishes direct and indirect speech and the structures associated with reporting verbs.
• Cambridge Dictionary, Reported speech: indirect speech — explains reported statements, questions, commands, backshift and deictic adjustments in English.
• British Council LearnEnglish, Reported speech: statements — practical English-language comparison of direct and indirect reporting.
• MQM Council, The MQM Error Typology — quality-assurance framework for classifying accuracy, addition, omission and other translation defects.

These references inform the terminology and English-specific examples. They do not imply that English tense and quotation rules are universal. For other language pairs, qualified linguistic, style and domain references must determine the natural realisation. The original exercises above are constructed teaching examples, not extracted testimony, and readers should never treat them as real quotations from named people.

14. Final principle: preserve the chain of responsibility

Translation succeeds when the reader can still determine what happened between speaker and listener. The sentence may gain different word order, a different tense choice, a different reporting construction or different punctuation. But the distinction between evidence and allegation, thought and speech, question and answer, request and command, observation and inference, speaker and narrator must remain legible. When that boundary is stable, translators can write naturally without compromising the people whose words they carry.

For the full translation workflow and the next specialised branch, return to the Master Art of Translation root. This chapter’s independent identifier is EDKSG-TRANS-MASTER-WORLD-1720. Its job is not to replace any existing vocabulary, English grammar, causality or evidence chapter, but to protect the specific chain through which another person’s words become a readable target-language report.