Master Art of Translation | The Speaker Attribution, Quotation and Reported-Speech Fidelity System — How Said, Claimed, Alleged, Denied, Quoted and Confirmed Preserve Who Said What

Accurate translation of quotes, reported speech and speaker attribution is not just a matter of choosing the right words. Readers searching for how to translate direct and indirect speech, how to translate quotations accurately, reported speech translation examples, or how to preserve who said what in another language face a harder question: whose words are these, who is reporting them, and how certain is the report? A sentence can be grammatically flawless yet attribute an accusation to the wrong person, turn an allegation into a fact, or make a hesitant witness sound confident. This guide gives a practical method for preventing that damage.

In professional translation, news translation, legal-document translation, interview translation, academic translation, subtitles and AI-assisted translation, the invisible work is to retain an attribution chain. A quoted person, a reporting writer, an interpreter, a transcript editor, a document author and the final translator may each occupy a different position in that chain. The words said, claimed, alleged, testified, suggested, denied, confirmed, according to, reportedly and quoted as saying are not interchangeable decorations. They affect evidential status, responsibility and sometimes the practical consequences of a text.

This chapter teaches a repeatable speaker-and-quotation fidelity system: identify every voice; distinguish the original utterance from a retelling; preserve direct versus indirect speech; keep allegation, denial and confirmation at their true strength; track translation and quotation provenance; and test the final rendering against plausible hostile or mistaken readings. Worked examples below use fictional situations and English-language meaning glosses so the principles apply to many language pairs without pretending that one language’s grammar is universal. The goal is to move the words into another language without moving responsibility to someone who never assumed it.

• Start here: record the source, speaker, reporter, audience, quotation status and evidential status before translating.
• The hidden error: fluent sentences can silently convert a report into an assertion or a paraphrase into a direct quote.
• The method: annotate claims, recover voice boundaries, draft naturally, verify attribution and run an independent reverse-reading test.
• The diagnostic: ask who a reasonable reader will think said each proposition, and which evidence supports that belief.
• Next route: return to the Master Art of Translation root for the larger translation system.

1. Search intent: what readers actually need

Most readers begin with a small and apparently grammatical problem: how to translate “he said that,” how to turn a quotation into reported speech, whether “claimed” means the same as “said,” or whether a translation can retain quotation marks. Those questions are practical, but they have layers. In a school exercise, the task might be grammatical transformation. In journalism, the same structure determines whether a speaker is accurately represented. In research, it decides whether a source is quoted or merely summarized. In a legal interview, it can separate the witness’s words from the recorder’s interpretation. A page that answers only the tense-shift exercise leaves the most important professional problem unsolved.

Searches about translating statements, witness testimony, meeting minutes, captions, subtitles and disputed quotations often converge on the same underlying challenge: readers need to know what the translated record actually establishes. Does the record show that a person uttered a phrase, that another person summarized their view, that an institution issued a statement, or only that someone believes an event occurred? The translation must carry those distinctions even when the target language favors a different order of words or requires an explicit subject absent in the source.

The useful outcome is an auditable decision process rather than a fixed list of equivalent reporting verbs. A professional can use that process with a bilingual text, a machine-generated draft, a live interpreter’s notes or a transcript of a conversation. A student can use it to spot why “she insisted” is a poor substitute for “she said.” An editor can use it to repair a headline that strips away “according to the report.” The common task is preserving the relationship between a person, a message and the evidence that the message was communicated.

2. Hidden problem: attribution can drift while meaning appears intact

Translation is often reviewed at sentence level. Reviewers check whether each visible noun, verb, figure and condition has a corresponding target expression. Attribution failure can survive this check because the error is relational rather than lexical. All the actors are present, all the quoted words are present, and every sentence reads smoothly; yet the quotation has been attached to the wrong person, or an unverified assertion has become the narrator’s own statement. Those shifts can be more consequential than an ordinary vocabulary mistake.

Consider the English gloss “The chair said the analyst believed the estimate was preliminary.” The chair reports the analyst’s belief. The analyst does not necessarily guarantee the estimate, and the narrator does not confirm it. “The chair confirmed the preliminary estimate” collapses two speakers, removes the belief marker and upgrades the evidence. A line-by-line dictionary audit might miss this because “chair,” “analyst,” “estimate” and “preliminary” remain recognizable. A relationship audit catches it immediately.

An attribution boundary also separates what a person actually said from what a writer infers. “She acknowledged a delay” is not necessarily “she admitted responsibility for the delay.” The second expression carries a stronger implication of culpability. “The witness could not recall” does not mean “the witness denied.” Silence is not agreement; a request for clarification is not a concession; a reference to an unnamed source is not first-hand verification. The translator’s role is to preserve these boundaries, not to settle facts that the source leaves unsettled.

3. The six layers of a reported statement

Build an attribution record with six layers. The first is the underlying event, which may or may not be known to have happened. The second is the proposition, the content asserted or discussed. The third is the original speaker, who made the relevant utterance or produced a statement. The fourth is the reporter or intermediary, who repeats, summarizes, records or interprets it. The fifth is the current document voice, which may endorse, distance itself from, or merely relay the report. The sixth is the translator’s intervention, which should transfer the earlier layers rather than become a seventh unmarked source of claims.

The same report can encode each layer differently. “A participant told the moderator that a supplier had notified her of an interruption” contains a participant, a moderator, a supplier and an interruption. Only the participant is directly described as telling the moderator; the supplier’s notice is reported by the participant. If a target sentence says “The supplier notified the moderator of the interruption,” it has changed the communication event itself. This is why a speaker map often needs arrows, not just highlighted names.

Keep the map small enough to use. For each clause record: who speaks; who is described as speaking; which source is direct; what proposition is carried; whether the proposition is affirmed, denied, questioned or alleged; and how the final reader knows that information. When a sentence contains nested speech, build nested entries rather than forcing it into a single owner. This prevents the habitual flattening that turns complicated accounts into simple but false narratives.

4. Direct speech, indirect speech and free indirect discourse

Direct speech presents an utterance as quoted words. Indirect speech reports its content through another narrator’s grammar. Free indirect discourse, especially in fiction, can blend a character’s perspective with a narrator’s surface language. The three modes are related but cannot be switched mechanically in every context. A direct quotation creates an expectation of fidelity to the quoted wording, subject to any clearly signaled translation or transcription policy. An indirect report allows grammatical and lexical restructuring while still requiring fidelity to content and stance.

Languages vary in tense backshifting, pronouns, demonstratives, verb forms and quotation conventions. A target-language indirect report may naturally change “I am leaving tomorrow” into a construction referring to the original speaker and the original next day. But not all languages apply English-style backshift, and the date of narration may matter. Preserve the time of the original speech before calculating the target expression. Do not confuse a classroom transformation rule with the evidential obligations of a translated document.

Free indirect discourse demands particular care. A novelist may write “Of course they had forgotten her. Why would anyone remember?” Although no speech tag appears, the perspective may belong to a character. Flattening the passage into a neutral authorial fact destroys irony and psychological distance. In a documentary or legal text, an untagged viewpoint is usually a reason to seek context, not an invitation to invent a speaker. Genre governs how the reader interprets voice.

5. Reporting verbs are decisions, not synonyms

Reporting verbs carry stance. “Said” often describes an utterance without overtly judging its truth. “Claimed” can imply that a proposition is asserted but unverified or contested, depending on context. “Alleged” commonly signals an accusation that has not been established; “confirmed” suggests an evidentially supported verification; “admitted” may imply acceptance of something adverse; and “denied” encodes rejection of a proposition. Those values are not identical across languages or institutional registers, but the distinctions must be examined before choosing a natural target equivalent.

A reporting verb also defines what the speaker is responsible for. “Promised” commits to future action; “predicted” describes an expectation; “warned” highlights anticipated risk; “recommended” proposes action; “ordered” asserts authority; “requested” does not necessarily. If translation replaces “asked” with “ordered,” it changes the social power of the event. If it replaces “suggested” with “confirmed,” it changes its evidential force. Treat the verb as a compact annotation of speech act, confidence, social relation and narrative distance.

Create a verb-function matrix for a project rather than a naive bilingual list. For each recurring source expression, record its normal function, the permissible target choices, disallowed near-synonyms, relevant genre and two contrasting examples. When uncertain, choose a neutral formulation that preserves known facts and flag the ambiguity if it changes stakes. The target need not mimic the source word-for-word, but it must not assign a new rhetorical act to the speaker.

6. Evidentiality: who knows, and how?

Evidentiality is the linguistic expression of information source. Some languages use dedicated grammatical resources to mark witnessed information, inference or reported evidence; other languages rely more on lexical choices such as “apparently,” “reportedly,” “according to” or “I heard.” Translators should not assume that the absence of an English-looking word means the absence of source-evidence marking. A small particle, suffix, verb form or construction may tell the reader whether a fact is observed, inferred, rumored or quoted.

Preserving evidentiality does not require making every implicit feature explicit. It requires understanding what the source actually communicates in its context and avoiding unwarranted upgrades. An evidence marker may be obligatory grammar with a broad interpretation, not a skeptical editorial comment. Conversely, removing a clear marker like “according to unnamed employees” can leave a translated headline claiming independent knowledge it does not possess. The right balance depends on language structure, discourse convention and target purpose.

The most useful evidentiality audit asks three questions. What does the source speaker actually know? What source of knowledge is grammatically or lexically indicated? What level of commitment will a target-language reader reasonably infer? If those answers diverge after translation, repair the sentence before editing for elegance. This audit belongs especially in reporting, scientific summaries, product claims, judicial material and public statements where evidential status matters.

7. A repeatable method: the VOICE record

The method in this guide is called VOICE: Verify the source, Outline the speaker chain, Identify the speech act, Carry the exact status, Examine the reader’s inference. It is deliberately compact. A translator can execute it in seconds for a routine message or document it in a separate sheet for a high-stakes record. The point is to reduce the cognitive load of repeatedly reconstructing who owns each utterance.

Verify the source means checking whether the words come from audio, an original-language transcript, a translated interview, a third-party summary, an official release or a screenshot. Outline means separating original speaker, quoting writer and other intermediaries. Identify means naming the act: assertion, hypothesis, allegation, denial, instruction, recommendation, question, concession, apology or commitment. Carry means selecting target-language syntax that preserves that act and its degree of certainty. Examine means reading the final sentence as if seeing it for the first time, then asking what it would lead a reasonable reader to believe.

Add an escalation rule. If the speaker is uncertain, the quotation has multiple plausible readings, a fragment has been detached from its original question, or the target language demands a distinction the source does not resolve, record the uncertainty and request context. When the issue cannot be resolved, use a transparent note or a carefully scoped translation instead of silently filling the gap. This is a method of fidelity, not a claim that every document can be made fully unambiguous.

8. Quotation boundaries, nested voices and punctuation

Quotation marks can be a visual cue to who is speaking, but quotation practice varies across languages, publishers and media. Some locales favor angle quotes, others typographic double or single quotes, and punctuation placement conventions differ. A consistent target style can change the shape of the marks without changing the number or ownership of quoted passages. What must remain visible is where the quoted voice begins, ends and contains another quoted voice.

Nested quotations require extra care. A journalist may quote an expert saying “The notice used the word ‘safe,’ but did not define it.” The expert is quoting a notice, and may be questioning its terminology. If the inner quote disappears, the sentence can imply that the expert personally declared the system safe. In the opposite direction, inserting quotation marks around a paraphrase may falsely present the words as an exact record. Treat changes to nesting as meaning-bearing edits, not only typography.

In multilingual web layouts, quotes may include right-to-left names, punctuation and left-to-right identifiers. Directionality errors can make a quote appear to belong to the wrong label. W3C internationalization guidance discusses isolated inline spans and direction handling; Unicode CLDR provides locale quotation data. These are presentation tools, not substitutes for source verification. A proper review checks both logical text order and actual rendered display.

9. The difference between translation, transcription and quotation

A transcription represents speech in written form under a chosen transcription policy. A translation represents meaning in another language. A quotation may reproduce wording from a source or from an already published translation. These operations overlap but are not interchangeable. A cleaned transcript can silently remove false starts, repetitions and uncertainty; a faithful translation of that cleaned transcript still cannot recover the omitted speech. A published English translation of an original-language statement is not proof that the speaker uttered the English words.

A credible record therefore states which layer is being quoted. “Translated from the interview in Language A” is different from “quoted from the published translation.” In academic writing, discipline-specific style rules may treat the writer’s own translation as a paraphrase rather than an exact-language quotation. In journalism, the handling of translated quotations requires an explicit editorial policy. In litigation, documentation of the source medium and interpreter may be critical. Do not assume one format answers all use cases.

Preserve the trace: recording or original document identifier, time marker or paragraph number, original language, transcription policy, translator or reviewer, version, quotation boundaries and any substantive editorial intervention. The trace does not make a claim true. It makes the steps from observed source to published target inspectable. When a source cannot be accessed, the translation should not imply it has been independently checked.

10. Interviews, depositions and question-answer context

Answers depend on questions. A witness’s “Yes” may affirm a narrow question, acknowledge hearing a statement, or agree only that an event is possible. Translating “Yes” into a full declarative sentence without preserving the question can turn a limited reply into a sweeping concession. Similarly, “I do not know” describes lack of knowledge; it need not deny that an event happened. The Q-and-A pairing is a unit of meaning.

Preserve interruptions, speaker changes and corrections according to the brief. When an interviewer asks several questions in one turn, identify which proposition an answer addresses. If the speaker corrects “Tuesday” to “Thursday,” do not smooth away the correction unless the delivery explicitly calls for an edited summary. A real-time interpreter may be producing interpreted speech, not an authorized transcript of the exact original utterance. Record each operation accurately.

Use an adjacency test: read the statement with the immediately preceding and following turns. This catches apparent contradictions that vanish when the question scope is restored. For high-stakes records, keep timestamps and stable speaker IDs. If a diarization tool guesses the speaker, mark the guess for human verification instead of turning it into a permanent attribution.

11. News, headlines and accusations

Headlines compress information and often drop function words. That compression is dangerous when a source merely alleges a fact. “Inquiry says supplier hid records” differs from “Supplier hid records”; “Workers say safety checks were skipped” differs from “Safety checks were skipped.” The first versions attribute claims. The latter appear to assert the underlying event directly. Translation must preserve that distinction even when a short target headline looks more attractive.

A second problem is selective quotation. Words can be translated accurately yet appear misleading when detached from a negation, exception or question. “I did not say the project was safe” cannot be shortened to “the project was safe” as a quote. A headline that quotes the most dramatic fragment of a nuanced answer should be reviewed for whether readers will infer a commitment the speaker did not make. If a quotation is translated from a secondary publication, that provenance belongs in the editorial record.

Do not manufacture balance by upgrading every person to an equal epistemic status, and do not weaken facts supported by the original source. Preserve who takes responsibility for a statement, what the evidence establishes and what remains disputed. The Associated Press principles explicitly emphasize attribution and context in reporting; their guidance is a useful editorial reference, not a universal translation code that replaces a project’s specific brief.

12. Research, science and academic quotation

Scientific claims travel through nested attribution: a paper reports experimental results; a review summarizes the paper; a news story quotes the review; a translated summary recasts the news. Each layer may add interpretation. “The authors observed an association” must not become “the researchers proved the cause” because the latter adds a causal conclusion. A translator should identify whether a reported finding is a direct result, a hypothesis, an interpretation, a limitation or another writer’s description.

Academic quotation also raises questions of source language and published translations. If the target reader sees a translated sentence inside quotation marks, identify whether it is a quotation from a published translation or your own rendering of original-language words. Follow the citation style governing the work. The key intellectual distinction is between reproducing a text, rendering its meaning and proposing an interpretation of that text.

An effective review preserves hedge words, sample limitations, confidence intervals, inferential verbs and disagreement. “May suggest,” “is consistent with,” “the authors argued” and “the report stated” occupy different evidential positions. Convert none into an unqualified result for smoothness. A bibliographic link or citation marker is not decorative; it lets readers trace the chain back to the primary source.

13. Technical logs, customer support and incident reports

Technical incident writing is full of attributed statements that are not confirmed system facts. A customer reports an outage; monitoring shows elevated latency; an engineer suspects a configuration change; a later review identifies a root cause. Translation should not collapse all four into “the system failed because of the change.” Each expression belongs to a different information state with different evidence.

Preserve the distinction between an event log and a person’s interpretation of that log. A message saying “User reports failed login” does not establish that authentication failed; a monitoring alert saying “possible data loss” does not establish data loss; and a closure note saying “issue appears resolved” may be provisional. Translators of support tickets and engineering postmortems must distinguish report, observation, diagnosis, action and verification.

Use stable field labels: reported by, observed by, confirmed by, time, evidence, status and open question. If a tool automatically summarizes an incident conversation, inspect whether it has attributed an explanation to the wrong engineer or erased the initial uncertainty. The target should let an on-call reader make the same operational decisions the source permits.

14. Subtitles, dubbing, live captions and conversation

Audiovisual translation works under timing and space constraints. Compression may be necessary, but the speaker’s identity and speech act often matter more than ancillary fillers. A character who says “I suppose he might have left” is not making the same commitment as “He left.” A subtitle that cuts the hedge changes the plot, while a dub that changes a request into a command may alter how an audience perceives a relationship.

Use visual and acoustic evidence carefully. Mouth movement, shot changes and a speaker’s voice can help with diarization, yet an off-screen speaker or overlapping dialogue can mislead. A transcription engine may assign speech to the loudest person rather than the actual speaker. The translator should check speaker turns against the original audio and relevant context whenever possible, and should mark unresolved identity in the production record.

Humor, sarcasm and irony complicate speech verbs. “Great job,” spoken after a mistake, may be critical rather than praising. A translation can reproduce the literal words while losing the character’s stance; conversely, adding “sarcastically” when the performance is ambiguous is an interpretive intervention. The right solution depends on audiovisual evidence and whether the format allows a note or only spoken language.

15. Literature, history and oral testimony

Literature uses voice as an artistic device. Narrators can be unreliable; a first-person account may contradict what the story reveals; a historical narrator may use labels that belong to the period rather than the translator’s present vocabulary. Changing a reporting verb can alter the moral alignment among character, narrator and reader. A literary translator must preserve purposeful uncertainty and shifts in narrative distance as well as propositional content.

Oral testimony often includes pauses, self-correction, emotion, imprecise chronology and reported conversations. A clean summary might be appropriate for a museum panel, but an archival transcript has different obligations. Do not silently convert a participant’s remembered dialogue into an exact contemporary quotation. “She remembered him saying” is different from “he said.” The former names the memory as a source.

Historical letters can contain quotations copied from other letters, newspapers or oral reports. Trace the immediate document and distinguish it from the alleged original utterance. A quotation’s repeated appearance in several records does not necessarily provide independent confirmation if those records copy one another. Translation should retain that source dependency rather than make a copied report appear independently corroborated.

16. AI and machine translation: fluent voice, misplaced responsibility

Machine translation and generative systems can produce fluent reported speech while misplacing the speaker, particularly when pronouns are ambiguous, turns are unlabeled or a paragraph includes nested quotations. A system may also replace a cautious verb with a more decisive one because the latter fits common prose patterns. Fluency is not evidence that the model has preserved attribution. The risk grows when an AI summary is treated as the source instead of a derivative artifact.

Work from the highest-quality original you are authorized to use. Keep protected data handling appropriate to the project. Mark speaker turns and quotation boundaries in the input where possible. Review the output against the original, not only against another machine-generated paraphrase. Use a structured quality taxonomy to classify errors: omission of a source marker, added certainty, wrong entity, altered speech act, unsupported quotation, or source-to-speaker inversion.

Do not use back-translation as the sole proof of correctness. A wrong target statement can back-translate into plausible English, especially if the same model family reproduces the same assumptions. Pair automated checks with focused human source comparison and test the most consequential sentences first. ISO 17100 covers professional translation-service processes, not raw machine-translation post-editing as a substitute for its specified workflow. State tool roles accurately rather than claiming automatic conformity.

17. The VOICE workflow in twelve operational steps

First, define the job and its acceptable intervention level. A school paraphrase, a literary translation, a news quotation and a legally sensitive record need different editorial controls. Second, obtain the best available source and note its language, format, version and provenance. Third, identify the document’s narrator and record each speaker turn. Fourth, divide compound sentences into propositions and identify the owner of each proposition. Fifth, mark direct quotations, indirect reports, summaries and unverified recollections.

Sixth, classify the speech act and evidential status of each proposition: observed, inferred, reported, contested, denied, admitted, promised, recommended or merely asked. Seventh, preserve modal force, negation, conditional scope, temporal orientation and source markers. Eighth, choose natural target-language grammar and quotation conventions without changing those relationships. Ninth, document ambiguities and seek source clarification when consequences warrant it. Tenth, compare the source and target claim graphs.

Eleventh, read only the target and identify what an uninformed reader would infer about responsibility and proof. Twelfth, sign off the exceptions and store the record needed to recheck future revisions. A high-stakes workflow should assign different roles for drafting and independent revision. A low-stakes message may use a shorter version, but skipping provenance entirely is rarely wise when the source already depends on hearsay or quotation.

These steps support, rather than replace, professional project specifications. ISO 11669:2024 addresses project planning and requirements, while ISO 17100:2015 describes core translation-service requirements. Neither should be reduced to a magic quality badge. For an implementation, specify roles, documents, acceptance criteria and escalation before the first difficult case arises.

18. The audit rubric: how to diagnose failure without guessing

Check speaker identity: each proposition should remain owned by the right person or institution. Check quotation status: wording represented as exact must have an appropriate source; paraphrases should not acquire false quotation marks. Check speech-act force: requests, promises, orders, apologies, allegations and denials must not be swapped for polished near-synonyms. Check evidential status: sourced, apparent, inferred and independently confirmed claims are distinct. Check nesting: the person repeating an utterance is not automatically the person who originated it.

Check time and deixis: pronouns, “here,” “today,” “tomorrow” and tense refer to the original speaking situation unless editorially reframed. Check fragments and context: a short quote must not contradict its longer context. Check negative scope: “not sure that,” “did not deny that” and “denied that” are different. Check reader inference: the best test is what readers actually take the target to mean. A stylistically elegant sentence can still fail.

Record severity by consequence. A misplaced comma in a quote may be cosmetic; quotation marks around fabricated wording can be critical. An omitted “according to” in a controversial allegation may be critical even when all the content words remain. The MQM community’s translation-quality typology distinguishes accuracy, mistranslation, omission and other issue families; project teams can adapt that structure to attribution-specific review. For every flagged issue, save the source span, target span, error type, severity, corrective decision and review owner.

19. The speaker-claim matrix readers can reuse

• Voice owner: who is responsible for the utterance; include unknown or institutional when the source does not identify a person.
• Carrier: audio, published transcript, interview notes, third-party story, email, edited release, remembered conversation or machine-generated output.
• Claim type: factual assertion, belief, suggestion, allegation, denial, request, instruction, prediction, warning, promise, admission or quotation of another voice.
• Evidence: witnessed, externally documented, reported by a named source, reported by an unnamed source, inferred, unverified or disputed.
• Target expectation: what a reasonable reader will conclude about the source’s knowledge, confidence and authority.
• Reviewer decision: accept, rephrase, annotate, verify externally, request context, or preserve uncertainty because no verified resolution exists.

A useful working sheet gives every proposition an identifier such as Q01 or R02. Link the source phrase to the target phrase, label each intermediary, and record whether the text is a direct quotation, a translation of a direct quotation, a paraphrase or a report of somebody else’s recollection. The identifiers are editing aids; they are not public claims of scientific certification. If a sentence crosses more than one voice boundary, duplicate the row for each embedded proposition rather than forcing several speakers into one cell.

20. Worked diagnostics — Journalism and public statements

News writers often receive a press statement, a response from a different party and a journalist’s contextual narration in a single paragraph. The danger is converting a reported statement into an independent finding or producing a headline that appears to endorse one source. Every file below tests a different editorial boundary. The English phrases are teaching glosses, not certified translations from a particular language.

Case 001 — The anonymous engineer

Source-side situation: An engineer who requested anonymity said the tests were incomplete. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The tests were incomplete, the agency confirmed.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “An anonymous engineer said that, in their view, testing had not been completed.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The source is unnamed and the engineer’s status is not equivalent to institutional confirmation. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check whether the source actually works for the agency and whether there is a public test record. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 002 — The press release

Source-side situation: The company wrote that its new battery could last up to twenty hours. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The battery lasts twenty hours.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “According to the company, the battery may last as long as twenty hours.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Marketing claims and upper bounds should not become independent performance guarantees. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve could and up to; check whether the number is a laboratory maximum. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 003 — The mayor’s answer

Source-side situation: Asked whether the plan was final, the mayor replied, ‘Not yet.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The mayor rejected the plan.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The mayor said the plan had not yet been finalized.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Not yet addresses finality, not support for the entire proposal. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Restore the question and check later revisions before writing a definitive headline. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 004 — The source’s silence

Source-side situation: The spokesperson declined to comment on the allegations. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The spokesperson denied the allegations.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The spokesperson offered no comment on the allegations.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Refusal to comment does not express the truth value of a proposition. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Review the recording or email to distinguish silence, refusal and a formal denial. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 005 — The conditional quote

Source-side situation: The director said, ‘If the data hold, this may change our recommendation.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The director announced a changed recommendation.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The director said the recommendation might change if the data remained valid.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The condition and modal possibility are part of what was said. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Underline if and may before drafting; reject versions that treat the future change as done. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 006 — The post-publication correction

Source-side situation: A reporter corrected an earlier quotation attributed to a minister. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The minister corrected his original statement.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The publication corrected its representation of the minister’s words.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The publication’s correction and the minister’s own correction are distinct acts. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Compare the correction notice, original recording and first published wording. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 007 — The disputed attribution

Source-side situation: Two agencies circulated a statement, but neither identified its author. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Agency A issued the statement.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “A statement circulated by two agencies did not identify its author.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Distribution does not establish authorship or endorsement by each distributor. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Look for the issuing authority or a signed provenance record. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 008 — The second-hand interview

Source-side situation: A newspaper reported that a witness told local media about a loud noise. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The newspaper heard the noise.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The newspaper relayed a witness’s account of a loud noise given to another outlet.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The witness and the intermediary publication are different sources. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep the source chain visible unless the witness was separately interviewed. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 009 — The opinion headline

Source-side situation: An analyst argued that the merger could reduce competition. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The merger will reduce competition.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “An analyst warned that the merger might reduce competition.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: An opinion is not an adjudicated economic outcome. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep analyst and might; separate this assessment from official regulatory findings. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 010 — The clipped sentence

Source-side situation: The speaker said, ‘I cannot confirm that the equipment is safe.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The speaker confirmed that the equipment is safe.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The speaker said they could not confirm the equipment’s safety.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Removing the initial negative reverses the speaker’s position. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Compare the complete clip and preserve the negation with its scope. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 011 — The written quotation

Source-side situation: An article quotes words from a policy paper, not from a live interview. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The official said this at the press conference.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The article quotes the official’s published policy paper.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A written utterance should not become a fictional spoken event. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Identify the document title and whether the named official authored or merely approved it. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 012 — The editorial interpretation

Source-side situation: A journalist wrote that the minister’s answer seemed evasive. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The minister admitted she was evasive.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The journalist characterized the minister’s answer as evasive.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The judgment is the journalist’s, not the minister’s self-description. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Locate the evaluative adjective’s owner before translating the sentence. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

20. Worked diagnostics — Interview, oral history and memory

Spoken testimony is situated. A person’s response can be constrained by a preceding question, another speaker’s interruption, the recording conditions and their own hesitation. The translator’s job is not to make the person sound consistently eloquent; it is to preserve the evidentially significant parts of what occurred within the agreed transcription policy.

Case 013 — The narrow yes

Source-side situation: Interviewer: ‘Did you see the door open?’ Witness: ‘Yes.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The witness saw the suspect enter.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The witness confirmed seeing the door open.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Affirmation of one observation does not establish who entered through the door. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep the question and answer adjacent and avoid filling missing persons. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 014 — The recalled sentence

Source-side situation: She recalled her grandmother saying the river had once been wider. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Her grandmother stated the river was wider.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “She remembered her grandmother describing the river as wider in the past.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A memory of reported speech is not a contemporary verbatim record. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Use remembered or recalled unless independent evidence supports an exact quotation. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 015 — The false start

Source-side situation: I—well, I thought the alarm was off, but perhaps I was wrong. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “I knew the alarm was off.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The speaker initially thought the alarm was off, then expressed uncertainty.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Disfluency can carry genuine self-correction, not merely stylistic clutter. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Identify the project policy for false starts and preserve the changed certainty. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 016 — The overlapping exchange

Source-side situation: Two people speak simultaneously and the transcript assigns ‘Stop’ to one. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The manager ordered everyone to stop.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “An overlapping voice says ‘Stop’; speaker identity requires verification.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Automatic diarization can invent authoritative identity in a chaotic exchange. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Compare channels, voice characteristics and time-coded audio before labeling. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 017 — The vague referent

Source-side situation: She said, ‘They told me it was canceled,’ without identifying they. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The organizers canceled it and told her.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “She said an unspecified group had told her it was canceled.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Expanding an unnamed pronoun into a named institution creates evidence. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Query who they refers to or preserve the unresolved referent. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 018 — The interpreted interview

Source-side situation: The interviewer hears an interpreter say, ‘He is worried,’ after a longer response. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The witness said exactly ‘I am worried.’” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The interpreter summarized the response as an expression of worry.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: An interpreter’s rendering may not reproduce the exact original words. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Label which text is original, interpreted and subsequently transcribed. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 019 — The two questions

Source-side situation: Interviewer: ‘Was anyone hurt? Was the lift working?’ Participant: ‘No.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Nobody was hurt and the lift failed.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The participant answered ‘No’ after two questions; its intended referent is unclear.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A single negative may answer only the final question or both. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Seek the full recording and avoid creating two unverified claims. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 020 — The retrospective timeline

Source-side situation: In 2026 she said she had believed, in 2021, that the site was safe. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “She says the site was safe in 2026.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “In 2026 she recalled having believed the site was safe in 2021.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The time of belief and the time of reporting are different. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Write both dates on a timeline before translating tense and deictic words. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 021 — The nod

Source-side situation: A transcript says ‘[nods]’ after the interviewer asks about receipt. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “She admitted she received the package.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The record notes a nod, without specifying a verbal admission.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A gesture can carry meaning but the transcript’s annotation is not a spoken sentence. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check video, gesture convention and whether a question was visible. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 022 — The inherited quote

Source-side situation: An oral historian repeats a line attributed to a distant ancestor. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The ancestor said these exact words.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The historian recounts a saying traditionally attributed to the ancestor.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Traditional attribution is not direct audio or written evidence. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep the oral-source chain and date of the recorded account. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 023 — The changed answer

Source-side situation: At first the witness said blue, then corrected the color to green. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The witness consistently identified green.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The witness first said blue and then corrected the description to green.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Corrections are part of the testimony, especially when reliability is evaluated. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve both answers or identify that a cleaned summary was authorized. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 024 — The emotional qualifier

Source-side situation: The participant said, ‘I felt as though nobody believed me.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Nobody believed the participant.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The participant described feeling disbelieved.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Subjective experience is not direct proof of everybody else’s belief. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain felt and as though rather than presenting perceptions as settled facts. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

20. Worked diagnostics — Law, governance and official records

Legal and administrative records have their own evidential rules and local terminology. The examples here are educational, not legal advice. The chief fidelity concern is avoiding an unearned change in the status of a claim, a declaration, a denial or a decision. Translate the record that exists, not the verdict a reader might expect.

Case 025 — The filed allegation

Source-side situation: The claimant alleges the contractor failed to disclose an inspection. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The contractor concealed the inspection.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The claimant alleges that the contractor did not disclose the inspection.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: An allegation in a pleading is not itself a judicial finding. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve the pleading source and check whether later findings exist in a separate document. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 026 — The contested motion

Source-side situation: The respondent disputes the proposed calculation. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The respondent proved the calculation false.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The respondent challenges the proposed calculation.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Dispute is a procedural position, not proof of mathematical error. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Separate the party’s objection from any adjudicator’s ruling. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 027 — The no-knowledge response

Source-side situation: The deponent said they did not know whether the contract was signed. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The deponent denied signing the contract.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The deponent said they lacked knowledge of whether it had been signed.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Ignorance of a fact and denial of one’s action are distinct. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep the interrogatory’s exact scope and respondent’s first-person position. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 028 — The official minute

Source-side situation: Minutes record that the committee noted the suggestion. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The committee adopted the suggestion.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The minutes show the committee acknowledged the suggestion.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Noting a proposal does not signify acceptance or authorization. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check the decision register for an actual vote or resolution. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 029 — The quoted statute

Source-side situation: A judgment quotes a phrase from an older law. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The judge introduced a new rule.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The judgment reproduces language from the earlier law.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A cited legal text and a judge’s own instruction have different authors. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain citation and mark where the quoted passage begins and ends. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 030 — The reserved decision

Source-side situation: The panel said it would consider the application later. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The panel approved the application.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The panel deferred consideration of the application.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A promise to consider is not a decision on the merits. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Review whether the source uses deliberate future modality. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 031 — The witness warning

Source-side situation: An officer told a witness that refusal might affect the process. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The officer threatened punishment for refusal.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The officer warned of a possible procedural consequence.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The target must not increase the alleged legal sanction beyond the stated one. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Identify what consequence was actually described without speculating about law. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 032 — The official clarification

Source-side situation: The office clarified that the deadline had not changed. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The office extended the deadline.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The office stated that the original deadline remained in place.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Clarification can reaffirm rather than revise an operative date. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check the actual date and distinguish announcement from rule. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 033 — The unsigned draft

Source-side situation: An unsigned memo proposes that the department issue a statement. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The department issued the statement.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “An unsigned draft proposes a future departmental statement.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Draft proposal, final approval and public issuance are different stages. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep draft status and look for evidence of adoption. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 034 — The officer’s quotation

Source-side situation: The complaint reproduces what the applicant says the officer told them. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The officer formally ordered the applicant to leave.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “According to the complaint, the applicant reported being told to leave.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: There are two nested levels of report before the reader reaches the officer. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Record claimant, alleged officer utterance and complaint as separate nodes. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 035 — The admission of receipt

Source-side situation: The office acknowledges receiving the document without accepting its contents. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The office accepted the document’s claims.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The office confirmed receipt, not substantive agreement.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Administrative receipt should not be mistranslated as agreement or approval. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain the limiting phrase attached to receipt. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 036 — The judicial summary

Source-side situation: The court summarized the prosecution’s case before evaluating it. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The court found all allegations proven.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The court described the prosecution’s allegations at that stage.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A neutral recital of a party’s case is not a finding of fact. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Locate the separate findings and orders before using found or held. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

20. Worked diagnostics — Medicine, science and public health

Clinical and scientific communication routinely contains uncertainty, patient reports, observational results and professional recommendations. This section teaches fidelity, not diagnosis or clinical instruction. Incorrect attribution can make a patient’s symptom sound like a physician’s finding or a tentative result sound like established causation.

Case 037 — The patient report

Source-side situation: The patient says the tablet caused dizziness. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The tablet caused dizziness.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The patient reports dizziness after taking the tablet and attributes it to the medication.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Temporal association and patient attribution do not prove causation. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve reported and distinguish adverse event from confirmed reaction. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 038 — The clinician’s hypothesis

Source-side situation: The clinician suspects a viral cause but has not ordered confirmation tests. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The clinician diagnosed a viral infection.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The clinician considers a viral cause possible, without confirmed testing.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Suspicion and diagnosis are distinct levels of certainty. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep suspects or possible and any missing test status. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 039 — The adverse event narrative

Source-side situation: The trial report records symptoms that participants described. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The trial confirmed the medicine causes all symptoms.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The report records participant-reported symptoms during the trial.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A report of events is not automatically a causal adverse drug reaction. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check whether investigators assessed relatedness separately. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 040 — The association claim

Source-side situation: Researchers observed an association between two variables. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Researchers proved one variable causes the other.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “Researchers reported an association, without establishing causation.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Observation, statistical association and causal inference are different propositions. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain the study design and any stated limitation. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 041 — The preliminary result

Source-side situation: The laboratory described the reading as preliminary. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The final laboratory result confirmed the condition.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The laboratory issued a preliminary reading.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A preliminary value can later be revised or contradicted. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check result state and authorizing signatory. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 042 — The participant’s memory

Source-side situation: A participant remembers being told to avoid a certain food. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The clinician instructed everyone to avoid it.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The participant recalls receiving an instruction about the food.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A patient’s recollection should not become a universal clinical rule. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve memory-source framing and avoid medical generalization. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 043 — The guideline recommendation

Source-side situation: A consensus group recommends additional assessment in selected cases. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The group mandates assessment for every patient.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The group recommends assessment for certain cases.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Recommendation and legal or clinical mandate have different force. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Identify the selected population and strength of recommendation. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 044 — The result interpretation

Source-side situation: A review says evidence may be compatible with the proposed explanation. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The evidence confirms the explanation.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The review says the evidence may support the proposed interpretation.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Compatibility does not establish unique proof. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve may and the review’s caveats. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 045 — The questionnaire item

Source-side situation: The form asks whether a symptom occurred in the past week. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The form states that the patient had the symptom.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The translated form asks whether the symptom occurred in the specified period.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A question must not be converted into a recorded positive response. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check polarity, reference period and response options. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 046 — The family account

Source-side situation: The note says the caregiver reports that the patient slept poorly. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The physician observed poor sleep.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The caregiver reports that the patient did not sleep well.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Caregiver report and clinician observation are distinct evidence channels. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve caregiver as the reporting source. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 047 — The investigator’s caution

Source-side situation: The investigator said the numbers were too small for firm conclusions. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The investigator concluded the treatment failed.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The investigator cautioned that the sample was too small for firm conclusions.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Lack of precision is not proof of failure. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep the stated limitation rather than substituting a verdict. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 048 — The evolving case

Source-side situation: The initial report was later amended after additional tests. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The initial report remains the final authority.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The report was revised when further test information became available.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The original statement and corrected record have different statuses. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Reference the effective version and revision history. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

20. Worked diagnostics — Workplace, HR and negotiations

Workplace translation carries consequences for discipline, consent, negotiation and interpersonal trust. Ordinary words such as agreed, noted, promised or expected may become operational commitments. This set tests who made a commitment and whether it was actually accepted.

Case 049 — The manager’s suggestion

Source-side situation: The supervisor suggested moving the meeting to Thursday. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The supervisor ordered a Thursday meeting.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The supervisor proposed moving the meeting to Thursday.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A suggestion does not necessarily override the existing schedule. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check whether someone accepted the proposal. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 050 — The tentative availability

Source-side situation: The colleague said they might be available after lunch. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The colleague confirmed attendance after lunch.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The colleague indicated possible availability after lunch.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Might is a capability or possibility marker, not a booking. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep the provisional modality. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 051 — The forwarded rumor

Source-side situation: An employee forwarded a message claiming layoffs were planned. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The employee announced layoffs.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The employee forwarded an unverified message alleging planned layoffs.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Forwarding text need not endorse it or prove it true. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Identify original sender and any employee commentary. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 052 — The meeting acknowledgment

Source-side situation: The participant wrote, ‘Noted,’ after receiving a plan. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The participant approved the plan.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The participant acknowledged receiving the plan.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Acknowledgment is not necessarily assent to the content. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Examine whether subsequent explicit approval exists. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 053 — The disputed promise

Source-side situation: One party says the other promised an extension; the other disputes it. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Both parties agreed to extend the deadline.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “One party alleges an extension was promised, but the claim is disputed.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Competing recollections should not be harmonized into consensus. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve both sources’ positions and date of each statement. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 054 — The prospective target

Source-side situation: The manager described a stretch target, not a guaranteed result. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The manager guaranteed the target.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The manager presented the figure as an aspirational target.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A target and a promise carry different obligations. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep words such as aim, hope or target where applicable. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 055 — The disciplinary allegation

Source-side situation: A letter says an employee is accused of breaching a rule. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The employee breached the rule.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The letter records an accusation of a rule breach.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Allegation and established finding require separate evidence. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check whether an investigation decision follows. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 056 — The personal objection

Source-side situation: The staff member said they were uncomfortable with the process. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The staff member refused to participate.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The staff member expressed discomfort about the process.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Emotion or concern does not necessarily amount to refusal. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Look for an explicit refusal or request for changes. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 057 — The draft settlement

Source-side situation: The parties exchanged a draft summary of proposed terms. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The parties signed a binding agreement.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The parties exchanged proposed terms in draft form.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A negotiation document is not the same as assent or execution. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve draft status and verify signatures if relevant. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 058 — The coordinator’s relay

Source-side situation: The coordinator said the supplier had promised to send samples. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The coordinator personally promised samples.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The coordinator reported the supplier’s promise.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The relay must not shift the obligation to the messenger. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Trace the promise to its original maker. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 059 — The risk warning

Source-side situation: The team warned that the timeline could slip. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The team announced a delay had occurred.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The team identified a risk of future schedule slippage.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Risk prediction and realized incident differ. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve could and the future orientation. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 060 — The conditional approval

Source-side situation: The reviewer said approval would follow if missing documents arrived. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The reviewer approved the application.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The reviewer described approval as conditional on receipt of missing documents.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A conditional future decision is not current approval. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep condition, actor and required evidence together. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

20. Worked diagnostics — Software products, incident tickets and AI summaries

Technology artifacts often mix user’s accounts, automated notifications, operators’ observations and later root-cause statements. A smooth translation can destroy the incident timeline or assign technical certainty to an unsourced report. Treat every status as an attributed observation until confirmation is documented.

Case 061 — The customer complaint

Source-side situation: The customer reports that sign-in sometimes fails. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The platform is permanently offline.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “A customer reports intermittent login failures.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: One user’s experience is not proof of global outage. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Compare the user report with uptime metrics. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 062 — The alerting system

Source-side situation: An automated alert says potential data loss detected. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The team confirmed data was lost.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “An automated alert flagged possible data loss.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: An alert is a trigger for investigation, not a verified conclusion. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve potential and identify the monitoring source. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 063 — The engineer’s suspicion

Source-side situation: An engineer suspects a cache mismatch. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Engineering confirmed the cache was the cause.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “An engineer identified a cache mismatch as a working hypothesis.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A hypothesis may be revised after debugging. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep the suspect label until incident review confirms a cause. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 064 — The release note

Source-side situation: Release notes say the update is intended to improve stability. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The update eliminated every stability problem.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The notes describe a stability improvement objective.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Intended outcome is not guaranteed achieved performance. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check whether test results separately support a claim. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 065 — The service message

Source-side situation: The status page says maintenance is scheduled for Friday. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The service was shut down on Friday.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The status page announces planned Friday maintenance.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A future scheduled action should not become historical execution. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep the schedule status and relevant timezone. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 066 — The automated summary

Source-side situation: An AI summary says the operator approved the rollback. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The operator definitely approved the rollback.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The AI-generated summary attributes approval to the operator; verify against the original chat.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Derivative AI output may hallucinate a speaker or decision. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Inspect the source thread and explicit approval message. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 067 — The support handoff

Source-side situation: A support agent relays that engineering requested more logs. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Support ordered the user to reset their device.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “Support reported engineering’s request for additional logs.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A request from engineering is not automatically the support agent’s personal instruction. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep request origin and requested item precise. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 068 — The bug report title

Source-side situation: A user titles a report ‘App deletes my files’. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The software is confirmed to delete files.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The user alleges that files disappeared after using the app.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Issue titles may describe perceptions rather than verified defects. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check reproductions and avoid presenting the report as a fixed cause. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 069 — The resolved appearance

Source-side situation: The incident note says service appears restored. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Service has been permanently repaired.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The incident note indicates service appears to have recovered.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Apparent restoration is a provisional observation. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain appears and any monitoring window. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 070 — The reproduced defect

Source-side situation: An engineer reproduces the login failure in a test environment. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Every customer is experiencing the failure.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The engineer reproduced the login failure under test conditions.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A confirmed reproduction establishes a defect under conditions, not prevalence. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Name environment and reproduction conditions. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 071 — The quoted user interface

Source-side situation: A support guide quotes the button label ‘Retry’. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The operator instructs a user to retry immediately.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The guide quotes the interface label ‘Retry’.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A label mentioned in documentation is not necessarily a command. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve quote boundaries and identify do-not-translate UI tokens. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 072 — The root cause statement

Source-side situation: A final report confirms a database lock caused the outage. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The initial customer identified the database lock.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The final investigation attributed the outage to a database lock.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Later confirmed analysis should not be retroactively assigned to the first reporter. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep event sequence, source of the finding and final status. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

20. Worked diagnostics — Education, schools and research interviews

Educational texts blend formal instructions, feedback, teacher interpretation and student accounts. A translation can turn feedback into discipline, a question into a command or a learner’s hypothesis into a proved conclusion. The objective is to preserve educational speech acts without reusing the protected language-learning owners’ broad curricular content.

Case 073 — The teacher’s observation

Source-side situation: The teacher observed that the student had become quieter. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The student admitted they had a problem.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The teacher noticed reduced participation.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Observed behavior and the student’s own explanation are separate. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep teacher as observer and do not diagnose motive. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 074 — The parent question

Source-side situation: A parent asks whether the assignment deadline can move. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The parent demanded a deadline extension.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The parent requested information about a possible extension.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: An inquiry is not necessarily a demand. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain interrogative force and level of politeness. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 075 — The rubric comment

Source-side situation: The reviewer wrote, ‘This claim needs evidence.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The reviewer said the whole essay was false.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The reviewer requested supporting evidence for the claim.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Local feedback on evidence is not a global verdict. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve the referent of this claim. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 076 — The student hypothesis

Source-side situation: A student suggested the change might be caused by temperature. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The student proved temperature caused the change.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The student proposed temperature as a possible cause.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A hypothesis should remain tentative. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep suggested and might. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 077 — The head’s announcement

Source-side situation: The principal said a consultation would be opened. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The principal announced a final policy.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The principal announced a forthcoming consultation.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Consultation and decision are different stages. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check whether a later policy has been issued. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 078 — The classroom quote

Source-side situation: A lesson quotes a historical speaker’s words in translation. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The original speaker used the modern target-language words.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The lesson presents a translated quotation attributed to the historical speaker.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Historical translation should not imply original-language identity. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Name the original work or source where possible. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 079 — The survey summary

Source-side situation: The survey team says students reported increased confidence. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The survey proves students objectively learned more.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The survey team reports higher self-rated confidence.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Self-report and tested performance are distinct outcomes. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep measurement instrument and reporter separate. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 080 — The feedback paraphrase

Source-side situation: A teacher paraphrases a learner’s concern to a guardian. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The learner used those exact words.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The teacher summarized the learner’s expressed concern.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Paraphrase and verbatim quotation are different records. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Do not add quotation marks unless source wording is available. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 081 — The provisional grade

Source-side situation: The portal labels an assessment score as provisional. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The final grade has been fixed.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The displayed score is provisional.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Provisional evaluation is not authoritative finalization. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve status and check the date of final review. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 082 — The appeal summary

Source-side situation: The student states that the instructions were unclear. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The institution accepted its instructions were unclear.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The student argues that the instructions lacked clarity.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A student’s contention does not establish an institutional concession. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep appellant and deciding body distinct. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 083 — The peer advice

Source-side situation: A classmate recommended reviewing the diagram. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The instructor required diagram review.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “A classmate suggested reviewing the diagram.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Advice from a peer has different authority than course instructions. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve speaker role and optionality. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 084 — The research consent

Source-side situation: An interviewer asked whether a participant wished to continue. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The participant agreed to continue.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The interviewer asked for the participant’s preference about continuing.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Consent is established by a response, not by a question. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep inquiry separate from any subsequent affirmative answer. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

20. Worked diagnostics — Finance, commerce and consumer communication

Consumer and financial material frequently presents forecasts, offers, market commentary and verified transactions side by side. A translator must preserve whether a number was announced, estimated, requested, received, verified or disputed, and by whom. This is an informational writing exercise, not financial advice.

Case 085 — The forecast

Source-side situation: Management forecasts revenue growth next quarter. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Revenue growth was recorded next quarter.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “Management expects revenue to grow in the following quarter.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A forecast is a prospective estimate, not an achieved result. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain forward-looking status and named management source. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 086 — The analyst estimate

Source-side situation: An analyst estimates the product could sell ten thousand units. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The seller guarantees ten thousand sales.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “An analyst estimates potential sales of about ten thousand units.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A third-party estimate is not a vendor guarantee. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve attribution, approximation and conditional possibility. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 087 — The buyer’s complaint

Source-side situation: A customer says the refund has not arrived. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The bank has confirmed withholding the refund.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The customer reports not yet receiving the refund.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Nonreceipt may have several causes; it is a customer report. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check transaction status before asserting misconduct. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 088 — The accepted offer

Source-side situation: The seller says a bid was received but not accepted. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The seller accepted the bid.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The seller acknowledged receiving the bid without accepting it.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Receipt and acceptance are different commercial acts. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve both statuses as recorded. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 089 — The unaudited figure

Source-side situation: A memo describes an unaudited internal revenue estimate. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The audited accounts confirm the number.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The memo contains an internal unaudited revenue estimate.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A source label defines evidential quality. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep estimate and unaudited visible. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 090 — The promotional quote

Source-side situation: A testimonial says, ‘This is the best service I’ve tried.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The business has been independently rated the best.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “A customer testimonial praises the service as their personal favorite.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Subjective endorsement is not a third-party ranking. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep testimonial source and subjective language. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 091 — The price proposal

Source-side situation: The vendor proposes reducing the fee if volumes rise. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The vendor reduced the fee.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The vendor offered a conditional fee reduction linked to higher volume.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: An offer has not necessarily taken effect. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain if, offered and future state. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 092 — The market rumor

Source-side situation: A trader reports hearing that a merger might be announced. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “A merger has been officially announced.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “A trader relays an unverified rumor about a possible announcement.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Hearsay about a future announcement has multiple uncertainty layers. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep rumor source and possibility separate. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 093 — The financial restatement

Source-side situation: The organization issued revised figures after finding an error. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The original figures remain authoritative.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The organization restated figures following an identified error.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A later corrected version differs from the original publication. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve version and correction attribution. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 094 — The speculative endorsement

Source-side situation: A commentator says the policy may increase prices. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The authority confirmed a price increase.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “A commentator predicts a possible price rise.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Commentary is not regulatory notice or observed effect. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain commentator and modal prediction. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 095 — The receipt of funds

Source-side situation: The cashier said the invoice was issued; no payment status was given. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The invoice was paid.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The cashier confirmed issuance of an invoice, not payment.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Invoice creation does not establish settlement. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check separately for a receipt or payment record. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 096 — The withdrawal notice

Source-side situation: The company says it intends to withdraw the offer. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The offer has already been withdrawn.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The company announced its intention to withdraw the offer.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Intention and completed withdrawal differ. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep future intention and effective timing. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

20. Worked diagnostics — Literature, dialogue and narrative viewpoint

In fiction, the attribution question includes who the narrator wants us to believe, what a character actually knows and whether irony is being signaled. Literary flexibility belongs to syntax and style; it does not permit casual destruction of perspective. These examples illustrate rhetorical mechanics rather than prescribe one universal literary solution.

Case 097 — The unreliable narrator

Source-side situation: The narrator boasts that everyone admired him while scenes show avoidance. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Everyone admired the narrator.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The narrator insists he was admired, despite contextual reasons to doubt him.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Narrator assertion may be intentionally unreliable. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep the boast’s ownership and avoid endorsing it in narrative voice. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 098 — The sarcastic compliment

Source-side situation: A character says, ‘Wonderful timing,’ as someone arrives late. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The character sincerely praised punctuality.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The character remarks ‘Wonderful timing’ in a context suggesting irony.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Tone comes from context, not the dictionary alone. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Consider performance cues before adding overt sarcasm labels. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 099 — The whispered rumor

Source-side situation: A villager whispers that the woods are haunted. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The narrator confirms ghosts inhabit the woods.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “A villager whispers a rumor that the woods may be haunted.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A story-world rumor should not become narrator-certified fact. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve whispered, rumor and source viewpoint. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 100 — The free indirect question

Source-side situation: The narration says, ‘Why had they left her behind?’ inside a character’s thought. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The author asks the reader for a factual explanation.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The passage retains the character’s inward question within the narration.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Perspective can shift without quotation marks. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Read surrounding focalization and pronouns before deciding speaker. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 101 — The misquoted poem

Source-side situation: A character misremembers a line from a poem. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The poem itself contains the character’s altered line.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The character imperfectly recalls a poetic line.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Misquotation can be an intentional characterization device. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Compare source poem and character’s version; preserve difference. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 102 — The conditional confession

Source-side situation: A character says, ‘Perhaps I knew more than I admitted.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The character confessed full knowledge.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The character tentatively suggests having concealed some knowledge.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Perhaps limits commitment; more is not all. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep degree and scope of self-accusation. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 103 — The quoted letter

Source-side situation: The story includes a letter inside a narrator’s account. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The narrator speaks the letter aloud.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The narrator reproduces a letter written by another character.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Written and spoken voices create different narrative layers. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain typographic or framing cues to the letter. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 104 — The implied accusation

Source-side situation: One character asks, ‘Were you there when it happened?’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The character accused another of causing the event.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The character asks whether the other person was present.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: An insinuation may exist, but the explicit act is a question. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve deliberate ambiguity instead of announcing a charge. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 105 — The rhetorical denial

Source-side situation: The character says, ‘Me, afraid? Never.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The character provided a verified psychological fact.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The character emphatically denies being afraid.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Dialogue expresses self-presentation, not necessarily truth. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Maintain rhetorical energy and voice ownership. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 106 — The deceptive quotation

Source-side situation: A villain quotes a hero accurately but omits the next qualifying sentence. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The hero endorsed the villain’s full interpretation.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The villain selectively quotes the hero, leaving out a qualification.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Selective quotation is part of the plot’s deception. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Restore narrative awareness of the omitted context where the text signals it. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 107 — The uncertain memory

Source-side situation: The narrator thinks, ‘Had she said leave or live?’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The narrator remembers precisely what was said.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The narrator cannot decide between two possible remembered words.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Uncertainty can be the entire narrative mechanism. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve alternatives without choosing one. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 108 — The narrator’s aside

Source-side situation: An omniscient narrator calls a plan foolish, while characters call it bold. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The characters admitted the plan was foolish.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The narrator evaluates the plan as foolish; the characters call it bold.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Separate authorial judgment from character vocabulary. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep contrasting valuations in their distinct voices. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

20. Worked diagnostics — Multilingual meetings, diplomacy and interpretation

A meeting may generate source-language speech, interpreter renderings, shorthand notes, drafted minutes and an approved summary. Each is an artifact with different authority. These cases teach how not to use a derivative representation to rewrite the commitments originally made.

Case 109 — The interpreter’s clarification

Source-side situation: An interpreter asks the chair to repeat a technical phrase. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The delegate requested a policy change.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The interpreter requested repetition for accuracy.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: An interpreter’s procedural intervention is not a delegate’s substantive position. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Label interpreter turn separately. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 110 — The delegate’s proposal

Source-side situation: The delegate proposed discussing a draft next week. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The delegation adopted the draft.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The delegate suggested placing the draft on a future agenda.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Discussion and adoption are different procedural stages. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Check agenda and decision records. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 111 — The chair’s summary

Source-side situation: The chair summarized a debate as ‘general support’ while objections remained. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “All delegates unanimously approved the proposal.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The chair characterized the discussion as generally supportive, despite recorded objections.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Chair summary does not cancel dissent. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep the qualifying phrase and dissent context. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 112 — The interpreted promise

Source-side situation: An interpreter renders a long speech as ‘We will help.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The speaker guaranteed unlimited assistance.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The interpreted message conveyed an intention to assist.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: An interpreter’s compression may omit scope and conditions. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Consult original audio before assigning contractual force. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 113 — The translation of a translation

Source-side situation: A press office translates an intermediary-language release. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The target is directly translated from the original speaker’s words.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The target derives from an intermediary-language version of the release.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Relay translation adds another layer where drift can occur. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Record both language stages and source documents. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 114 — The adopted minute

Source-side situation: Draft minutes say a motion passed, but approved minutes amend that record. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The unapproved draft is final evidence.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The approved minutes provide the adopted record, with the earlier draft documented as superseded.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Record status and attribution intersect. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Identify which version the task requires; never silently blend them. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 115 — The noncommittal reply

Source-side situation: A delegate says, ‘We will take the proposal under advisement.’ The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The delegation accepted the proposal.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The delegate said the proposal would receive consideration.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Taking under advisement is not assent. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve procedural noncommitment. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 116 — The off-record remark

Source-side situation: A note marks a participant’s remark as off the record. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The public statement quotes the remark as authorized.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The notes record an off-record remark subject to agreed restrictions.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Translation does not erase access constraints or transform authorization. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Confirm publishing permissions before reproduction. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 117 — The joint communiqué

Source-side situation: The document states a joint position approved by signatories. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “One particular delegate personally authored every sentence.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The communiqué records a collectively approved institutional position.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Corporate authorship and individual speech are different. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep institutional source and approval scope. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 118 — The tentative consensus

Source-side situation: A facilitator notes emerging consensus, not a vote. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The committee formally voted for the measure.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The facilitator observed signs of developing consensus.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Consensus assessment is not a completed recorded vote. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve tentative status and reporting role. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 119 — The question to the chair

Source-side situation: A member asks whether an item can be postponed. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The member formally moved postponement.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The member inquired about the possibility of postponement.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Question, procedural motion and decision must remain distinct. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Use formal meeting vocabulary only when the source warrants it. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 120 — The translated quotation for a communiqué

Source-side situation: A draft public note quotes an interpreter’s English rendering. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The delegate uttered those English words.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The draft quotes the interpreted English rendering of the delegate’s statement.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The immediate quote source is the interpreted text. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Trace original speech and approvals before publishing exact quotation marks. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

20. Worked diagnostics — Digital platforms, user-generated content and social media

Posts, screenshots, reposts and edited clips travel rapidly, often without intact provenance. Social attribution is especially vulnerable to mistaken endorsement. Quoting a claim to criticize it is not the same as adopting it, and a reposted image does not automatically reveal its author.

Case 121 — The repost without caption

Source-side situation: A user reposts a disputed statistic without writing a comment. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The user independently calculated and asserted the figure.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The user redistributed a post containing a disputed statistic.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Reposting and originating a claim are not synonymous. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Retain repost status and avoid inventing intent. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 122 — The debunking quote

Source-side situation: A fact-checker quotes a false rumor before refuting it. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The fact-checker endorsed the rumor.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The fact-checker reproduced the rumor as the claim being checked.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: The surrounding refutation determines stance. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve the quotation and explicit debunking context. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 123 — The screenshot claim

Source-side situation: A screenshot appears to show an official message, but authenticity is unverified. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The agency definitely issued the message.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “An unverified screenshot purports to show an agency message.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A visual appearance is not independent authentication. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Mark authenticity status and search official source where authorized. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 124 — The deleted post

Source-side situation: A story says a post was deleted, without identifying who removed it. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The platform censored the post.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The post was removed or became unavailable; the remover is unspecified.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Deletion event and agent identity are different claims. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Do not assign motive or actor without source. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 125 — The translated hashtag

Source-side situation: A hashtag repeats a slogan used by protesters. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “Every participant formally endorsed the full manifesto.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The hashtag reproduces a slogan associated with the campaign.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A slogan can be a quotation, label or expression of support. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Distinguish use, quotation and endorsement in context. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 126 — The verified badge

Source-side situation: An account shows a platform badge; the article quotes its message. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The named individual personally authored the message.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “A message was posted from the indicated account.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Account status alone may not prove who typed the message. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Use the actual source’s identification policy. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 127 — The edited clip

Source-side situation: A short clip removes a speaker’s initial ‘I disagree that’. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The speaker backed the following proposal.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The full statement rejects the proposal; the short clip is misleading.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Negation outside a clip changes attributed position. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Compare long-form original and timecode. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

Case 128 — The automated subtitles

Source-side situation: Social captions attribute speech to the person visible on screen. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The visible person necessarily spoke the sentence.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “Automatic captions assign the utterance to a speaker that needs verification.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: On-screen appearance is not diarization evidence by itself. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Review original audio and speaker turn. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: For reviewer calibration, invite two competent bilingual readers to annotate the source independently, before they see the proposed target. Compare their assignments of speaker, certainty and quotation status. Disagreement identifies a genuine source-reading question; agreement followed by a divergent target identifies a translation problem.

Case 129 — The emoji response

Source-side situation: Someone replies with a thumbs-up after a complicated proposal. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “They legally accepted every condition.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The person responded with a thumbs-up emoji.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Emoji interpretation depends on context and agreement conventions. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Avoid adding a formal acceptance verb without support. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: In a production workflow, attach a short exception note only when it adds information a later editor will need: source unavailable, unclear pronoun, competing transcript, a nested quotation, or required approval. This is not an invitation to clutter every sentence with caveats. It is a way to preserve the evidence when a consequential decision could otherwise be lost.

Case 130 — The quoted headline

Source-side situation: A post says ‘Newspaper: prices set to rise’ and criticizes that headline. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The poster forecast higher prices.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The poster quoted a newspaper headline while discussing it critically.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: A mentioned claim belongs to the quoted publication. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Preserve nested quote and critical stance. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful counterfactual test is to replace the source reporting verb with a stronger neighbor—say with confirm, suggest with announce, recall with witness—and see whether the proposition changes. If the target already implies the stronger expression, it has likely overtranslated the original. The repair is usually small, but its effect on responsibility is large.

Case 131 — The parody account

Source-side situation: A parody account posts a fictional statement attributed to a celebrity. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The celebrity issued the statement.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “A parody account published a fictional statement about the celebrity.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Genre and source status shape attribution. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Keep parody framing where the source provides it. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: A useful transfer exercise is to recast the same example as a short headline, a formal record and a casual message. Change register and length, but preserve the same proposition owner. If compression forces a choice, keep the attribution marker before optional descriptive detail. The result should remain intelligible even when the reader sees only the short version.

Case 132 — The multi-source thread

Source-side situation: A long thread quotes one expert, then paraphrases a second without transition. The source is presented as a fictional English meaning gloss, not as a sworn transcript or a verified quotation from a real person. The translator’s first responsibility is to identify the actual carrier of this information and decide whether the statement is direct speech, an indirect report, a quoted document, a summary or an inference. The distinction determines what the target version can responsibly claim.

Misleading target reading: “The entire thread is the first expert’s opinion.” This phrasing is dangerous because a reader could assign a different speaker, degree of certainty, procedural status or practical commitment to the underlying proposition. A translation may sound confident and idiomatic while still moving a person’s responsibility or an evidence marker across a boundary it was meant to preserve.

Faithful English back-gloss of the target choice: “The thread contains a direct quotation from one expert and a separate paraphrase of another.” The wording is an illustration of the required meaning, not a compulsory expression in every target language. In an actual assignment, use the target language’s natural grammar, punctuation, reporting constructions and register. What cannot be optional is the relationship between the source’s speaker, proposition, speech act and level of evidence.

Why the distinction matters here: Formatting loss can merge two sources into one. Review the sentence as a miniature claim graph. Ask who made the original utterance, who is now relaying it, whether the target voice appears to endorse it, and which words limit or strengthen its force. If the target changes any of these without authority from the source, classify the change as an attribution or accuracy failure, not merely a stylistic preference.

Diagnostic and repair: Add clear attribution markers without fabricating direct quotes. Then run two independent readings. In the source-to-target reading, highlight evidence markers, time reference, negation, speaker labels and the exact boundary of any quoted words. In the target-only reading, ask a reviewer unfamiliar with the source to identify who said what and how firmly the statement was made. If that reviewer would tell a materially different story, revise the target or disclose the unresolved ambiguity.

Transfer drill: To test robustness, imagine that the target is later quoted by a third person. Would its wording still reveal which part is fact, which part is a report and which part is judgment? If not, adjust the attribution phrase now. Misunderstanding often arrives not in the first translation but in the next reuse of that translation.

21. Production laboratories: twenty complete review exercises

Laboratory 1 — Back-translation trap

Translate a cautious indirect statement into a target language, then back into English. Explain why returning to similar words does not prove speaker ownership survived. Compare the underlying claim graph instead. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 2 — Direct-to-indirect shift

Rewrite a direct quotation in indirect form. Check the pronouns, tense, time of original speech and whether the writer now appears to endorse the utterance. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 3 — Nested allegation

Design a three-speaker report: a witness tells an editor that an unnamed person alleged a defect. Preserve the three sources without implying that the defect is confirmed. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 4 — Denial with scope

Translate a sentence that denies a particular time but not the entire event. Ask a reviewer which proposition remains open after reading the target. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 5 — Corrected transcript

Compare a raw transcript, edited transcript and translated excerpt. Mark which changes are transcription interventions rather than translation decisions. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 6 — Hearsay marker

Choose a source construction that grammaticalizes reported evidence. Identify how your target language can carry it without inventing skepticism. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 7 — Anonymous quotation

Translate a source’s assertion from an unnamed person without replacing anonymity with a fictional job title or institutional endorsement. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 8 — Institutional statement

Compare a named spokesperson’s personal remark with an agency’s signed statement. Distinguish the legally or editorially responsible source. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 9 — Screen quotation

Embed a right-to-left person name in a left-to-right quotation card. Test how labels, quotation marks and neutral punctuation render. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 10 — Headline compression

Reduce a three-sentence account to twelve words while preserving allegation or conditional status. Explain which facts were safe to omit. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 11 — AI summary audit

Give an automated model a multi-speaker exchange and inspect whether it invents a conclusion, changes a request into an order or omits a source marker. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 12 — Chronology of reports

Separate when an event allegedly happened, when a source described it and when the translation was published. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 13 — Interview answer

Make an answer meaningfully depend on the preceding question; test what happens if the question is removed. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 14 — Scientific hedge

Translate three related phrases: suggests an association, supports an explanation, demonstrates a cause. Do not collapse them into one level of confidence. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 15 — Fictional unreliability

Write a narrator who boasts and a scene that contradicts the boast. Translate without changing the deliberate literary distance. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 16 — Customer service report

Distinguish a user claim, service telemetry, engineer hypothesis and final incident review. Check that no statement gains premature confirmation. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 17 — Policy approval

Make a committee note that merely acknowledges a recommendation. Produce a target version that does not read like formal adoption. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 18 — Quoted translation

Quote a published translation and separately render the original into your own language. Explain which text the quotation marks are reproducing. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 19 — Reputational allegation

Translate a contested accusation with appropriate source markers and right of reply; keep both positions distinct rather than arbitrarily smoothing conflict. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

Laboratory 20 — Final reviewer challenge

Hide the source from a reviewer, show the target alone and ask them to list the speakers and the evidential status of each proposition. Compare that list with a source annotation. The deliverable is not merely a fluent sentence. Produce a five-column worksheet with the source proposition, voice owner, speech-act label, evidence status and target formulation. Make a second reader explain what they would infer without access to the source. A successful laboratory documents any ambiguity that cannot safely be resolved from available evidence.

For a demanding version of this exercise, supply the same material in two formats: a formal, fully attributed report and a short reader-facing summary. Compare what each version makes explicit. A shortened version is acceptable only when it remains clear whether the writer is reporting someone else’s words or stating a finding independently. Write down exactly which omitted details were not needed and which must be restored to prevent a misleading inference.

22. Common failure modes and how to repair them

Attribution laundering

Failure signature: A statement originally made by a secondary source appears as an established fact in the final translation. Repair principle: Reinsert the reporting source and preserve whether it merely quoted, summarized or endorsed the claim. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Quotation inflation

Failure signature: A paraphrase is placed inside quotation marks and presented as the exact words of the original speaker. Repair principle: Remove false quotation marks or locate and accurately translate the original wording, noting the translation source. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Evidential upgrade

Failure signature: Words expressing rumor, possibility or observation are replaced by certainty and confirmation. Repair principle: Restore evidence markers and check whether corroboration is genuinely in the source. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Evidential downgrade

Failure signature: A clearly documented, confirmed finding is weakened into mere rumor for stylistic balance. Repair principle: Preserve the established source status rather than editorially changing the strength of evidence. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Actor replacement

Failure signature: An intermediary who passed on a message becomes the author of its content. Repair principle: Reconstruct the sender-receiver chain and assign each action to the correct participant. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Nested-quote collapse

Failure signature: A quotation inside another quotation loses its inner boundary and appears to be the outer speaker’s opinion. Repair principle: Mark nested spans, quotation levels and referents before rewriting punctuation. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Scope drift

Failure signature: A denial of one narrow claim becomes a denial of an entire event or allegation. Repair principle: Identify the exact proposition negated and express its boundaries in the target. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Deixis error

Failure signature: Tomorrow, here and I are interpreted from the time and place of the translator rather than the source speaker. Repair principle: Anchor speech time, venue and participants before choosing target-language forms. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Cross-medium confusion

Failure signature: A written letter is described as something said aloud at a meeting. Repair principle: Keep whether content was written, spoken, recorded, remembered or interpreted. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

False institutional authority

Failure signature: An individual’s speculation becomes an agency or organization’s confirmed position. Repair principle: Keep individual job title separate from formal authorization to speak for an institution. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Transcription smoothing

Failure signature: Disfluency, repair or qualification is silently removed, changing the speaker’s apparent certainty. Repair principle: Apply a disclosed transcription policy and preserve changes that affect evidence or intention. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Headline endorsement

Failure signature: Removing according to or an allegation marker turns a headline into a direct assertion. Repair principle: Retain the source in a natural concise form, even at the expense of a shorter headline. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Answer isolation

Failure signature: Yes or no is presented without the question that gives it propositional meaning. Repair principle: Restore the question or paraphrase only the explicitly supported proposition. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Diarization hallucination

Failure signature: A software-generated speaker label is treated as verified fact. Repair principle: Check audio/video and label uncertain identity until verified. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Secondary-source erasure

Failure signature: A translation based on another translation is described as direct from the original language. Repair principle: Record each source step and avoid claiming original-language access the editor did not have. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Narrator-character merger

Failure signature: Literary viewpoint is flattened into neutral narrator assertion. Repair principle: Preserve focalization, irony and character-bound perspective. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

False admission

Failure signature: Acknowledged receipt or awareness becomes admitted guilt or responsibility. Repair principle: Identify what was actually acknowledged and avoid importing adverse implication. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Future-to-present shift

Failure signature: Intention, promise, plan or proposed action is translated as a completed event. Repair principle: Restore aspect, timing and conditional status. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

Quotation-context loss

Failure signature: A fragment is accurate in isolation but reverses the meaning of the longer sentence. Repair principle: Read surrounding sentences and the source question before publishing the quote. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

AI fluency bias

Failure signature: A highly natural target is accepted without checking its source-attribution graph. Repair principle: Review high-risk speech acts against the original and record substantive corrections. A reviewer should document the exact source span and the corrected target span. Make a separate judgment about severity: a cosmetic difference in punctuation is not equivalent to replacing an allegation with a proven fact. Where a dispute could arise, preserve the evidence needed for later independent review.

23. Frequently asked questions about translating quotations and reported speech

How do I translate reported speech correctly?

Identify the original speaker, the proposition being reported, the reporting verb’s force, the source time and the evidence marker. Build natural indirect syntax in the target language, then check whether a reader would still assign the content to the same person. English tense backshift is not a universal rule for all languages.

Should quotation marks stay the same across languages?

Preserve direct-versus-indirect status and nesting, not necessarily the exact graphic shape. Apply the target locale’s conventions where appropriate, using an approved style guide. Do not put quotation marks around a paraphrase unless the publication’s explicit convention and source evidence justify that presentation.

Is ‘claimed’ the same as ‘said’?

Not always. Claimed may suggest an assertion whose correctness is unresolved or contested, while said can be comparatively neutral. Choose the expression that preserves the original stance. When the original does not clearly evaluate truth, avoid importing a suspicious tone merely to sound journalistic.

Is ‘alleged’ the same as ‘confirmed’?

No. Alleged presents a claim or accusation, generally without establishing it; confirmed signals verification or an authoritative affirmation. Converting one into the other changes evidential status. Check who is actually confirming what and on what basis.

What if the source language has evidential verb endings?

Analyze the grammatical form and its local discourse use. Determine whether it signals a report, inference, perception or another information source. Look for a natural way to preserve the distinction in the target, potentially using syntax or lexical markers. Do not mechanically add reportedly to every sentence without examining its effect.

Can I use indirect speech instead of a direct quotation?

Sometimes, depending on purpose and editorial permissions. Indirect speech allows restructuring but must preserve the speaker’s stance, conditions and essential content. In a record requiring verbatim representation, conversion may not be allowed. Check the brief and identify any intervention.

Can I quote my own translation of an original-language text?

This depends on citation and editorial style. Distinguish an original-language quotation, a quotation from a published translation and your own English rendering. Some academic styles treat a writer’s own translation as a paraphrase. Label the source and follow the applicable discipline’s instructions.

How should I translate an unattributed quotation?

Do not invent an author. Reproduce the statement with its unknown status or explain that the source gives no attribution when that matters. Seek the original source if the quote is central to the claim. Lack of source information is an uncertainty to manage, not an invitation to fill it.

Can translation fix a misleading original quote?

The translator may identify and flag a misleading fragment, but should not silently rewrite the source’s editorial position. Escalate to the commissioning editor, provide a contextual note where permitted, and separate faithful translation from independent fact-checking or editorial correction.

How do I translate yes and no in interviews?

Keep the preceding question and any clarifications. A bare yes can affirm only one component of a complex question. Do not expand it into a comprehensive factual declaration without support. Where the response is ambiguous, mark it and seek recording context.

How do I distinguish ‘denied’ from ‘did not admit’?

Denied actively rejects a proposition; did not admit describes absence of an admission, which may include silence, lack of knowledge or a different answer. These are not equivalent. Preserve the precise scope of what the source records.

Should I preserve hesitation?

Follow the transcription and delivery brief. Hesitation that changes a speaker’s apparent confidence, uncertainty or self-correction can be substantive. Unimportant fillers may be edited for readability in an authorized summary, but should not disappear from a forensic transcript without policy.

What about interpreting rather than written translation?

Interpreting is a separate service with real-time constraints and its own standards. The attribution principles still matter, but an interpreter’s words are not automatically an exact transcript of original speech. State the medium and identify which record the written translation derives from.

What is a source-of-source chain?

It is the succession of people or artifacts that carried a proposition: original speaker, interview, reporter, translation, newsletter and final article, for example. Each step can introduce compression or error. Record it when exact attribution, evidential strength or publication responsibility matters.

Why do subtitles often lose reported speech?

Subtitles face timing and reading constraints. Compression can erase words like apparently, according to or perhaps. These short markers can be essential to story and responsibility. Prioritize them when they determine what a character claims or knows.

How can an editor check AI-translated quotations?

Use source-aligned review. Mark speakers and quotation spans before comparing outputs, inspect negation and confidence markers, confirm named entities, and ask a fresh reviewer what the target implies. Avoid treating a back-translation from the same model as proof.

What if multiple people speak at once?

Use available audio channels, timecoded recording and speaker-labeling conventions. Where attribution remains uncertain, label it uncertain rather than guessing. In a production workflow, obtain approval for the chosen representation before publishing an attributed direct quote.

Can I make a quote more grammatical?

Only within the applicable policy. A readable edited quotation, faithful transcript and exact archival record have different obligations. Changing grammar can alter register, identity or meaning. Document significant intervention and consider paraphrase instead of pretending edited words are exact.

What is an attribution graph?

A small map of propositions and who says, reports or quotes them. It is useful because nested reported speech may contain several different voices in one grammatical sentence. Arrows make visible whether the target mistakenly attaches a proposition to an intermediary.

How do I handle pronouns in indirect speech?

Find the original referent for every I, you, he, she, they, we, here and this. Then express the same reference naturally in the target language. If the source pronoun is genuinely ambiguous, do not silently select an identity solely because the target sentence wants clarity.

Do different languages use different quotation marks?

Yes. Typographic conventions vary by language and locale, including preferred opening, closing and alternate marks. Unicode CLDR documents locale information; published style guides may have specific requirements. The semantic boundary of quoted material matters more than preserving an English mark mechanically.

Why does an accusation need special attention?

An accusation can affect reputation and legal exposure. The difference between a claim made by one party and an established finding must remain explicit. Preserve source, status, denials and available responses without inventing a verdict or weakening actual documented findings.

What is the difference between ‘reported’ and ‘observed’?

Reported typically identifies information communicated by a person or document; observed indicates what an observer or instrument registered. One individual can report an observation, but a reader should not assume the current writer witnessed it. Maintain the chain when it changes certainty.

How do I manage quotes in multilingual HTML?

Preserve semantic quote structure, correct language metadata where available, and isolate bidirectional fragments appropriately. Check actual rendered punctuation and names as well as underlying text order. W3C’s internationalization guidance describes direction handling for embedded phrases.

Is a signed statement the same as a spoken interview?

No. The signed text is a documentary artifact with its own authorship and approval status. A spoken interview is an utterance captured under different conditions. A publication can quote either, but it should identify which source is used.

Can a report be inaccurate without mistranslating any individual word?

Yes. The wrong attachment of a clause, omitted attribution marker or a misplaced quote boundary can alter the proposition’s apparent source. Review clause relationships and reader inference, not only terminology matches.

How do I mark a secondary translation?

State that the text was translated from an intermediate published version when relevant. Identify the languages, source documents and versions in the editorial record. Do not claim original-language comparison unless it occurred.

What if one source contradicts another?

Translate each position faithfully with its speaker and source status. Reconciliation, adjudication or fact-checking is a separate editorial task. If the target reader needs context, provide it according to the brief rather than secretly making the sources agree.

Where can I learn the broader grammar and vocabulary?

The protected Vocabulary Learning Hub owns vocabulary acquisition and word knowledge. How English Works owns English grammar, discourse and intra-language explanation. This child focuses on cross-language transfer of attributed voices, quoted words and evidence relationships.

What is the fastest final quality check?

Read the target alone and answer: Who said this? Did that person actually utter these words or are they paraphrased? Is the statement confirmed, denied, alleged or conditional? Then compare those answers with the source. If any material answer changes, the translation needs another pass.

24. Research references and scope of authority

The following references were consulted for terminology, standards scope, editorial attribution and multilingual presentation. Publication dates refer to the authorities’ documents, not to a claim that every project has independently demonstrated conformity. Standards and editorial manuals have distinct jobs: guidance about process, quality criteria, published quotation practice, source provenance and presentation should not be collapsed into one rulebook.

• ISO 17100:2015 — Translation services, requirements: The published standard describes core translation-service processes and resources; its stated scope excludes raw machine-translation post-editing and interpreting services.
• ISO 11669:2024 — Translation projects, general guidance: The March 2024 edition guides project specifications, needs and risk analysis, communication and workflow, rather than purporting to assess every translation output directly.
• W3C Internationalization Tag Set 2.0: The provenance data category identifies translation and revision agents; other data categories cover localization context and document linkage.
• MQM Community Group — multidimensional quality metrics: The error hierarchy provides practical vocabulary for accuracy, mistranslation, omission, addition and related quality review.
• Associated Press — News values and principles: Telling the Story: Attribution and quotation context are explicit elements of editorial integrity in the agency’s published principles.
• United Nations Editorial Manual — Quotations: The manual distinguishes quotations, omissions, accurate reproduction and related citation expectations for its publication context.
• Unicode CLDR charts and quotation locale data: Locale data can inform quotation typography, which must be kept distinct from semantic quotation ownership.
• W3C — Inline markup and bidirectional text in HTML: The technical guidance explains ways to isolate embedded text direction where multilingual quotations mix scripts.
• Royal Roads University — citing translated sources in APA style: The university’s October 2025 guidance distinguishes quoting published translations from one’s own translation in APA-based scholarly writing.
• Taboada — Reported speech and gender in the news: The September 2024 research demonstrates how speaker selection and reporting verbs affect journalistic representation.

25. Transfer, responsibilities and the next reading routes

This chapter is one specialist mechanism within the Master Art of Translation | The Complete System for Moving Meaning Between Languages. The root article owns the general translation architecture: source analysis, meaning invariants, terminology, human revision, technology and quality assurance. This chapter owns a narrower task: preserving who uttered, reported, inferred, denied, promised or confirmed each proposition while crossing language boundaries. Return to the root when you need the whole system rather than a single attribution mechanism.

When a translation error comes from insufficient word knowledge, collocations or lexical choice, continue to the Vocabulary Mastery Learning Hub rather than duplicating its vocabulary-learning curriculum. When the reader needs grammar, English reported speech patterns, tense, pronoun reference or discourse analysis inside English, continue to How English Works | Translation and Meaning Across Languages. Those established language-learning owners remain separate; the present guide does not replace or revise them.

Related specialist routes in this series include the Illegibility, Uncertainty and Editorial-Intervention Fidelity System for unreadable or inferred wording; the Original, Copy, Extract and Transcription Fidelity System for source-record status; and the Complaint, Objection and Appeal Fidelity System for the lifecycle of contested decisions. Each resolves a neighboring but distinct question.

The final discipline is simple to state and demanding to practice: a translator may change the language, sentence order, quotation typography and even the amount of explanation, but must never silently change who owns the assertion. Reliable translation carries not only what was said, but who said it, how it reached the record, what kind of speech act it was and how much that record really establishes. Preserve those relationships and a multilingual reader can still decide for themselves what the evidence means.