Master Art of Translation | The Pragmatic Equivalence System — How Speech Acts, Politeness, Implicature and Social Intent Keep Their Meaning

How do you translate politeness, requests, refusals, implied meaning and social intention accurately between languages? The difficult part of pragmatic translation is not always vocabulary or grammar. A sentence that looks like a question may be a request, a courteous statement may be a refusal, and a brief acknowledgement may not signal agreement. Reliable speech-act translation preserves what speakers are doing to one another, not merely the words they pronounce.

Readers searching for pragmatic equivalence in translation, how to translate indirect speech acts, politeness strategies across languages, conversational implicature, or cross-cultural translation accuracy face a hidden problem: language changes the social relationship between participants. An apparently fluent target sentence can redistribute authority, soften a prohibition, manufacture consent, erase a sincere apology or make a cautious claim sound definitive. A dictionary does not supply the speaker’s relationship to the listener, the stage of an exchange or the consequences of a mistaken response.

This guide explains a repeatable system for preserving communicative force, register, social distance, facework, stance, turn-taking and the boundaries of what the source actually supports. It uses original examples, practical diagnostics and counterfactual tests that apply across language pairs, including those where neither source nor target is English. It distinguishes source evidence from translator inference, offers a method for handling unresolved ambiguity, and connects to the wider Master Art of Translation architecture without duplicating its established vocabulary and English-language branches.

Architecture owner: EDKSG-TRANS-MASTER-WORLD-1230. This chapter owns cross-language preservation of interpersonal action and inferred intent, not stand-alone language learning, translation quality certification, or the previous chapter’s general reader-comprehension testing.

The answer first: translate the social act, then verify the words

The practical question is: after reading the translation, what would a reasonable member of the target audience believe they are being asked, permitted, warned, promised, refused or invited to do? Translation fails when the answer changes without authorization. This does not mean imitating the source language’s politeness formula word for word. It means preserving the source’s relevant obligations, options, uncertainty, relationship cues and intended action while using credible resources of the target language.

A workable model distinguishes six objects: the literal proposition, the speech act, the interpersonal relationship, the implied but recoverable message, the procedural consequence and the available evidence. The first object is what words state. The second is what the speaker is doing. The third determines which expression is socially fitting. The fourth is what a participant may infer. The fifth tells us why a mistake matters. The sixth prevents us from turning guesses about intention into facts.

The most important safeguard is the distinction between an interpretation and an invented translation. When a source permits several readings, a translator may need to preserve that ambiguity, ask a question or mark an unresolved choice. They should not quietly select the reading that is easiest to express. Where action or safety depends on the ambiguity, escalation and source clarification are part of accurate translation rather than signs of failure.

The hidden problem: words are only one layer of the transaction

A phrase such as ‘Could you review this?’ can be a request even though its grammar resembles a question about ability. ‘That might be difficult’ can be a refusal, a warning about feasibility, or merely a forecast. ‘Fine’ can accept a proposal or indicate reluctant compliance. None of those meanings resides in the word alone. They emerge from roles, prior turns, timing, medium, status, common ground and the shared conventions of a community.

The translator encounters a double uncertainty: what did the original speaker mean in this particular interaction, and which target-language construction will let a target audience infer the same socially important content? The first cannot be solved by fluency in the target language; the second cannot be solved by interpreting the source correctly. The two questions must be answered separately and then tested together.

The major loss is often invisible. Numeric mistranslation can be found by comparing digits. A mistranslated refusal may contain no conspicuous error, because all the nouns and verbs are correct. The target listener simply hears willingness where the source listener would have understood reluctance. Without an explicit pragmatic check, this can pass proofreading, bilingual alignment and even ordinary fluent reading.

How the mechanism works: twenty-four meaning channels

1. Locution versus action

A sentence can assert a fact while functioning as a complaint, warning or invitation. Its grammatical mood does not determine its full action. The key is to treat locution versus action as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Check the prior turn and the response the utterance makes relevant. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Avoid classifying every question as information-seeking and every imperative as coercive. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Ask an informed reader what response would count as appropriate. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

2. Directness

Languages differ in their conventional methods for encoding requests and refusals; directness also varies by genre, urgency and relationship. The key is to treat directness as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Preserve the degree of negotiability and permission rather than copying the source’s surface syntax. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Do not assume that a direct sentence is automatically rude or an indirect one automatically kind. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Contrast the target with an explicit order and with a purely optional suggestion. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

3. Politeness as relation management

Politeness signals not only warmth but social distance, entitlement, formality, respect and the wish to avoid imposition. The key is to treat politeness as relation management as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Identify speaker and audience roles, then choose a form that is appropriate to that specific relationship. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Avoid national-character stereotypes and unexamined rules about entire languages. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Test the same target line under a peer relationship and a hierarchy reversal. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

4. Facework

Participants manage dignity, autonomy and solidarity through mitigators, honorifics, justifications and acknowledgements. The key is to treat facework as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Preserve the relevant respect or independence without promising something the source did not promise. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Over-apologizing may concede responsibility; under-mitigating may turn a refusal into a rebuke. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Ask whether each participant retains the same room to decline or disagree. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

5. Implicature

Conversation can communicate something beyond what is explicitly asserted, based on relevance, cooperative expectations and shared knowledge. The key is to treat implicature as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Carry recoverable implication when needed, but separate strong contextual inference from mere speculation. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Do not replace every hint with a categorical declaration, particularly in negotiations. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Explain the implied reading and cite the source cues that justify it. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

6. Presupposition

Questions and statements can take previous events, identities and responsibilities for granted without expressly establishing them. The key is to treat presupposition as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Protect what is assumed and what is disputed; do not transform a loaded question into an admitted fact. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

A neat target sentence can accidentally validate an unproven premise. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Ask whether a cautious reader would now believe the premise was verified. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

7. Epistemic stance

Words such as may, apparently, reportedly and likely calibrate how strongly a speaker commits to a claim. The key is to treat epistemic stance as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Translate the strength and source of knowledge, not only the claim’s main noun and verb. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Strongening ‘might’ into ‘will’ can create an invented guarantee. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Rate source and target commitment as certain, supported, tentative or reported. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

8. Deontic force

Must, should, may, can and prohibited encode obligation, advice, permission and possibility; grammatical equivalents can be deceptive. The key is to treat deontic force as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Test whether a target reader believes they have an option or a duty. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Do not reduce a mandatory condition to a helpful recommendation. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Ask the reader whether noncompliance is allowed and what follows. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

9. Social roles and address

Some languages encode familiarity, kinship, institutional rank and age relationship through pronouns or address forms. The key is to treat social roles and address as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Record only relationship facts actually known and use appropriate forms under a confirmed brief. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Assuming rank, gender or intimacy can misrepresent identities and social distance. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Run a role-reversal test with an informed target-language reviewer. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

10. Turn-taking

The place of an utterance in a conversation changes its function; an answer, invitation, challenge or correction may use the same words. The key is to treat turn-taking as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Translate against the turn sequence, not as an isolated sentence from a spreadsheet. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Segment-by-segment translation can misclassify a short reply. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Read the preceding and following turns and identify the expected next move. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

11. Preference organization

Acceptances and refusals often differ in timing, prefaces, explanations and indirectness. The key is to treat preference organization as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Preserve a meaningful delay or account where it signals reluctance, while keeping subtitles and captions manageable. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Removing all hesitation can change a reluctant no into an eager yes. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Ask a reader whether the exchange sounds like acceptance, reluctance or refusal. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

12. Interactional repair

Clarifications, self-corrections and acknowledgements signal that common understanding is being reconstructed. The key is to treat interactional repair as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Make it clear who corrected what and whether the previous statement was withdrawn. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Smoothing every repair into a single perfect sentence can erase a dispute. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Check whether the final target still records the original misunderstanding. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

13. Discourse markers

Well, actually, so, anyway and comparable forms organize turns, resistance, inference and shifts in topic. The key is to treat discourse markers as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Translate the discourse function rather than searching for a mandatory one-word pair. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Omitting ‘actually’ can erase correction; forcing a marker can invent disagreement. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Ask what conversational transition a target reader hears. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

14. Register

A social action changes when delivered as a memo, direct message, ceremonial remark or support interaction. The key is to treat register as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Preserve the level of formality appropriate to the setting and the source’s purpose. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Do not treat relaxed phrasing as evidence of consent or formal language as evidence of authority. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Read target lines in the real channel where they will appear. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

15. Voice and ownership

Pronouns, passive structures and group labels distribute responsibility for decisions and actions. The key is to treat voice and ownership as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Preserve who performs, authorizes or promises each action; distinguish singular from collective commitments. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Replacing ‘I’ with ‘we’ can create institutional backing that does not exist. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Trace every action to an actor and compare the source. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

16. Quotation and reported speech

Reported remarks are not the same as statements the document itself endorses. The key is to treat quotation and reported speech as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Maintain attribution, quote boundaries and the speaker’s level of commitment. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

An unsignalled paraphrase can turn criticism cited for discussion into an official position. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Ask who would be held responsible for the claim in each version. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

17. Irony and sarcasm

Literal praise may carry negative evaluation when tone or context makes that interpretation compelling. The key is to treat irony and sarcasm as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Preserve the interactional signal only when there is adequate evidence and genre permission. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

A translator can easily invent sarcasm where the source is sincere. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Compare cues from intonation, sequence, punctuation and participant reactions. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

18. Humour and play

Jokes, teasing and mock-formality depend on shared knowledge and mutually recognized boundaries. The key is to treat humour and play as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Decide whether to preserve the specific image, conversational relief or character relation. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Replacing a joke with unrelated mockery can damage character and credibility. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Ask whether the target audience understands the playful frame without new harm. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

19. Negation and polarity

Negative questions and double negatives can affect both factual scope and interactional expectations. The key is to treat negation and polarity as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Check the answer the utterance requests and the proposition it negates. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Polite negative questions can be rendered as accusations by changing their force. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Test a yes and a no reply to see whether they map consistently. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

20. Temporal obligation

Words such as soon, later, by Friday and whenever possible set different expectations for timing. The key is to treat temporal obligation as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Keep hard deadlines hard, flexible windows flexible and urgency justified. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

A translation that makes ‘soon’ a fixed time invents a commitment. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Ask what is late, optional and due under each version. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

21. Prosody and pacing

Stress, pauses, pitch and speed can contribute to refusal, uncertainty and emotion, particularly in live interpreting. The key is to treat prosody and pacing as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

If audio exists, include relevant delivery cues without pretending they are universal codes. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Written transcripts can conceal hesitation or overinterpret pauses. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Use audio with an informed reviewer and compare alternative plausible readings. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

22. Platform affordances

A button, notification or recorded message may be short but still grant permission, demand action or imply completed status. The key is to treat platform affordances as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Preserve the action consequences and interface state, not just a concise phrase. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Friendly wording must not become a false assurance that something has happened. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Test the phrase at the point where users decide what to do next. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

23. Cultural and institutional scripts

Some familiar formulas carry different expectations about hospitality, apologies, deference or reciprocity. The key is to treat cultural and institutional scripts as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Use community and genre evidence rather than abstract beliefs about a culture. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

A literal customary formula may sound like an enforceable promise in translation. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Ask local reviewers what action they expect from the formula in context. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

24. Translator uncertainty

Not every pragmatic choice can be resolved from text alone; ambiguity is sometimes intentional. The key is to treat translator uncertainty as a testable layer of meaning. Begin with a literal reading, then ask what communicative action a participant would infer in the real situation. This prevents an apparently correct bilingual pair from winning merely because the vocabulary sounds familiar.

Keep an uncertainty register and ask targeted questions before the target is released. Write down the evidence: who speaks, who receives it, which event is already underway, what response is expected, and what remains uncertain. Separating those observations from the translator’s own judgement makes the decision reviewable and gives a second translator an opportunity to challenge it.

Confidently guessing recipient relationships or hidden promises is not faithful translation. The safeguard is not a vague instruction to ‘sound natural’. It is a comparison of practical consequences. Document the evidence, unresolved branch and owner of the decision. If plausible target readers disagree on the meaning, investigate whether the ambiguity belongs to the original or was introduced by translation. Preserve source ambiguity where possible; do not use fluent prose to conceal a weak inference.

The repeatable translation method: eighteen gates from source to release

Gate 1: Collect the full exchange

Recover the previous question, subsequent reaction, medium, speaker identity and purpose; one isolated segment is not a sufficient conversational unit. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Request the two turns on either side or record that context is unavailable. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: The reader identifies the wrong conversational action despite flawless grammar. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 2: Name the observable act

Mark whether the utterance requests, invites, declines, warns, reports, apologizes, proposes, corrects, jokes or remains genuinely undecidable. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Use two plausible readings when the record does not settle one. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: The draft attributes a definite demand to an ambiguous sentence. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 3: Separate fact from intention

Record literal propositions separately from inferred motives, emotional attitudes and interpersonal effects. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Support any inference with timing, wording, genre or explicit context. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: The target asserts that a speaker is angry when no evidence establishes anger. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 4: Map roles without stereotypes

Document confirmed status, familiarity, public or private audience and degree of institutional representation. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Do not infer gender, rank or closeness from a first name or national label. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: Role-dependent target pronouns or honorifics are chosen without evidence. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 5: Assess authority and obligation

Identify which actions are mandatory, optional, advisable, forbidden or conditional. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Build an obligation table with source wording and expected recipient action. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: A recommendation changes into a command or a prohibition changes into advice. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 6: Assess commitment and evidence

Record whether the speaker promises an outcome, estimates a probability, reports a source or merely states hope. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Keep commitments and evidential stance in different columns. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: The translation says ‘will’ when the source only says ‘may’. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 7: Identify source ambiguity

Mark unresolved referents, uncertain relationship cues, polyfunctional phrases and unclear refusals. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Ask a narrow question before publication or preserve the ambiguity in the target. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: The target’s precision is greater than the available source evidence. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 8: Choose target speech-act resources

Use a native and genre-appropriate request, refusal, warning or acknowledgement construction. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Compare at least two candidates that differ in directness or social stance. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: One fluent candidate sounds like an order while the other permits refusal. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 9: Rebuild turn structure

Keep the sequence of proposal, hesitation, refusal, explanation and next action when this sequence matters. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Avoid removing a meaningful correction simply to make the conversation elegant. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: The target exchange seems resolved even though the source remained disputed. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 10: Control facework

Decide which acknowledgement, mitigation, deference or solidarity is necessary for equivalent reception. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Check whether added apologies or compliments change responsibility. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: The source speaker seems to confess error only in the target version. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 11: Check lexical triggers

Audit modal verbs, discourse markers, polarity, tense, address terms and conversational formulas. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Use term resources for lexical options but decide by function. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: A word-level match survives while its implied duty does not. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 12: Check quoted and attributed claims

Maintain the boundary between what a speaker says, what someone reports and what the author endorses. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Track quotation marks, reporting verbs and evidence qualifiers. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: An opinion from one speaker becomes the platform’s official statement. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 13: Use a counterfactual relationship test

Ask how the line would sound with a different hierarchy, degree of intimacy or institutional role. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Where the result changes, confirm whether the source supplied that relationship. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: The target assumes superiority or familiarity that was never stated. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 14: Use a response-prediction test

Ask readers what they would do immediately after seeing the target line. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Compare predicted actions with the source’s intended actions, not just word meanings. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: Readers comply where the source intended optional participation. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 15: Perform two independent readings

First compare source and target for semantic and pragmatic fidelity; then read the target alone for credible interaction. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Have separate reviewers record issues before discussing them. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: A bilingual reviewer accepts a literalism that a target-only reader finds insulting. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 16: Prioritize by consequence

Assign greater review attention to consent, refusal, promises, deadlines, safety warnings and legal responsibilities. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Defer low-impact stylistic preference when core action meaning is incorrect. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: A beautifully phrased UI label gives an inaccurate status or permission. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 17: Record evidence and exceptions

Store source segment, version, chosen interpretation, alternatives rejected, review owner and unresolved questions. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Document which assumptions were confirmed and which were deliberately left open. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: A later source revision cannot reconstruct why a politeness form was chosen. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Gate 18: Release and observe

Test in the real audience channel, verify expected responses, and re-open errors when context changes. At this gate, write a concrete artifact rather than checking a box from memory: a context note, an act label, an evidence-strength rating, a role matrix, a candidate comparison or a reviewer decision. The purpose is traceability. Another translator should be able to see which property of the source was protected and why a target expression was selected.

Working action: Use feedback as evidence rather than as proof that every reader has the same convention. Ask the reviewer to state both a proposed reading and the observable cue that supports it. An explanation such as ‘this sounds nicer’ cannot establish equivalence. A useful explanation names a source constraint, a target convention and the behavior that the target expression is likely to invite.

Gate failure: A release repeatedly produces misunderstandings that reviewers dismiss as user mistakes. Return to the source rather than polishing the same candidate repeatedly. Where the error concerns a promise, responsibility, safety condition or consent, treat it as a meaning defect even if its grammar is perfect. Close the gate only after the corrected target is retested in the relevant interaction.

Failure modes: why fluent translations lose interpersonal meaning

Failure 1: Question mistaken for mere ability

An indirect request becomes a factual query; recipients may answer ‘yes’ and take no action. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Add turn and task context, then use an idiomatic request form that preserves optionality. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 2: Soft refusal rendered as acceptance

A diplomatic ‘that is difficult’ is heard as a challenge to solve rather than a decline. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Confirm intention with a source owner; where clear, select a target refusal with equivalent facework. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 3: Apology inflated into admission

A sympathetic ‘sorry to hear’ becomes responsibility for an event. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Distinguish empathy, regret, responsibility and remedy before choosing apology formulas. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 4: Advice inflated into an instruction

‘You may want to’ becomes mandatory direction. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Match the reader’s actual freedom to ignore it. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 5: Warning diminished into suggestion

‘Do not enter’ becomes ‘consider staying out’. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Keep negative obligation and hazard visibility. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 6: Tentative evidence asserted as fact

‘Reportedly’ or ‘appears’ disappears. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Preserve source attribution and degree of certainty. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 7: Ambiguous pronoun assigned identity

Gender, hierarchy or closeness is guessed from an absent context. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Use a neutral option when possible or request evidence for obligatory distinctions. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 8: Politeness formula translated literally

A customary exchange creates an artificial invitation, obligation or promise. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Use target equivalents grounded in participant expectations. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 9: Humour converted to hostility

Playful teasing becomes a literal insult. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Check the play frame and participant consent; mark uncertain cases. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 10: Sarcasm manufactured

Sincere praise is misread as irony. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Require context, not intuition alone, before reversing evaluation. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 11: Correction smoothed away

The target implies everyone agreed from the start. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Preserve repair moves and who revised the statement. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 12: Consent inferred from silence

No response is represented as acceptance. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Keep silence distinct from informed approval unless a verified rule establishes otherwise. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 13: Quoted claim promoted to endorsement

An allegation gains the narrator’s authority. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Protect reporting verbs and quotation boundaries. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 14: Deadline made more precise than source

‘Soon’ becomes a numerical day or time. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Preserve flexibility or query the sender. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 15: Institutional ‘we’ made personal

The target changes who is making a promise. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Track speaker accountability and collective voice. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 16: Literal ‘please’ over-generalized

A target directive becomes overly deferential or unusually aggressive. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Use genre and action force to select a natural form. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 17: False equality of formal titles

A source relationship is reduced to a title without a counterpart. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Preserve role information but resist inventing status markers. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 18: Translated emoji carries altered sentiment

A visual cue changes the perceived warmth or seriousness of the message. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Test the complete line in its communication platform. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 19: Turn delay removed

Reluctance or a refusal precursor disappears. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Retain the interpersonal function of pacing within medium constraints. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 20: Cultural maxim substituted carelessly

A familiar proverb changes the moral implication or stance. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Write the intended proposition and effect before considering substitutions. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 21: Confidence created by AI fluency

A plausible generated sentence masks missing relationship data. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Run independent target-user judgement and mark assumptions. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 22: Speaker intent over-claimed

The translator confidently states a motive only inferred weakly. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Document evidence levels, preserve ambiguity and escalate high-stakes decisions. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 23: Release context ignored

A correct line in a document becomes misleading on a confirmation button. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Test within actual UI or conversation and record expected user action. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Failure 24: Reviewer preference treated as proof

A stylistic preference erases another community’s valid convention. This is a pragmatic defect, not merely a question of taste. The translator should diagnose the source utterance’s propositional content and its social action separately, then decide whether the target creates a new obligation, an unsupported claim or a misleading emotional stance.

Distinguish hard meaning defects, evidence-based usability defects and acceptable alternatives. A controlled review asks two target readers to describe what action they would take and what relationship they infer. If their responses diverge, compare that divergence with the source audience’s likely range of readings. Correct newly introduced ambiguity, while leaving intentional ambiguity intact when the brief permits it.

Worked diagnostic laboratories: 96 real interaction problems

Every laboratory below presents an original English-language source example and a concrete translation decision. English is used as a visible notation for analysis, not as the required source or destination of translation. The suggested repair describes the function that a competent target-language expression must realize. Actual target-language wording depends on the language pair, relationship and genre; the guide does not pretend that one universal polite phrase works everywhere.

Diagnostic domain: Requests, questions and instructions

This domain turns requests, questions and instructions into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 1: Could you send the file before lunch?

Source and situated reading. “Could you send the file before lunch?” A collaborator needs the file to complete a review, not a report on the recipient’s technical ability. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Treating it only as an ability question invites ‘yes, I can’ without delivery. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A time-bounded request, softened by question form, whose deadline remains meaningful. Use a target-language request construction suited to collaborators, keeping the file and time expectation explicit. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask the reader what must happen and by when. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 2: Do you have a moment?

Source and situated reading. “Do you have a moment?” A speaker uses an opening before introducing an inconvenient request. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Publishing it as a request for an exact unit of time distorts the conversational opening. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A check of availability that prepares, but does not yet specify, a second request. Preserve the preparatory nature rather than inventing the later task. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the listener expects a follow-up. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 3: Please keep this confidential.

Source and situated reading. “Please keep this confidential.” A professional shares sensitive but non-classified discussion notes. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Treating ‘please’ as evidence the instruction is optional weakens confidentiality. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A request or direction to restrict disclosure, with seriousness set by context. Retain the restriction while using an appropriate form of courtesy. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether sharing it onward is permitted. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 4: Would you mind lowering the volume?

Source and situated reading. “Would you mind lowering the volume?” A neighbor needs a quieter environment and uses negative-question politeness. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Literal polarity can produce an inverted response or an accusatory tone. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A non-hostile request to reduce sound, not an inquiry into annoyance. Choose an idiomatic request pattern and check negative-question reply conventions. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask what change the recipient would make. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 5: It would help if the table were updated.

Source and situated reading. “It would help if the table were updated.” A reviewer comments on a draft during a team meeting. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Leaving it as a hypothetical observation may conceal the requested revision. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An indirect suggestion or request, not proof that completion is mandatory. Preserve the suggestion’s strength and identify who can authorize changes. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask if the recipient considers the edit required or proposed. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 6: Let’s begin with the unresolved question.

Source and situated reading. “Let’s begin with the unresolved question.” A facilitator redirects a discussion that has become repetitive. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A literal invitation may sound like a personal preference with no chairing force. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A procedural instruction from someone currently managing the agenda. Use facilitation language that begins the new topic without inventing disciplinary authority. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask participants what topic comes next. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Refusals, limits and boundaries

This domain turns refusals, limits and boundaries into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 7: That might be difficult for me.

Source and situated reading. “That might be difficult for me.” A colleague replies to an invitation with visible reluctance. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Translating it as a technical obstacle can invite pressure to solve the obstacle. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A likely soft refusal, though final intent needs the full exchange. Retain hesitation and do not assert a hard no without supporting context. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the listener expects participation. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 8: I appreciate the offer, but I cannot accept.

Source and situated reading. “I appreciate the offer, but I cannot accept.” A person declines a costly gift. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Removing the appreciation makes the rejection unnecessarily confrontational. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Gratitude followed by a clear boundary. Preserve both recognition of goodwill and nonacceptance. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the gift should be kept or returned. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 9: Perhaps another time.

Source and situated reading. “Perhaps another time.” A speaker ends a conversation after being invited to meet. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. An overly literal future promise can create an actual appointment expectation. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A polite deferral that may also function as a decline. Avoid introducing a definite future meeting. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether a date is now arranged. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 10: I would rather not discuss that.

Source and situated reading. “I would rather not discuss that.” Someone declines a personal question in a public chat. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Rewriting it as an inability implies the problem might be fixed by permission. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An autonomous boundary on topic, not a request for help. Select a firm but proportionate boundary expression. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether a follow-up question is invited. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 11: I do not think I can commit to Friday.

Source and situated reading. “I do not think I can commit to Friday.” A contractor responds to a proposed delivery date. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. The phrase can be wrongly softened into an assurance of best efforts. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Noncommitment to the deadline, without a confirmed alternative date. Maintain absence of guarantee and seek a feasible date. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether delivery on Friday is promised. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 12: No, thank you; I have enough.

Source and situated reading. “No, thank you; I have enough.” A guest refuses a second serving. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Deleting the courtesy makes a routine refusal sound like criticism. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A complete refusal plus a socially positive acknowledgement. Keep the refusal unambiguous and the social tone unthreatening. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether another serving should be offered. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Apologies, responsibility and empathy

This domain turns apologies, responsibility and empathy into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 13: I’m sorry this happened.

Source and situated reading. “I’m sorry this happened.” A support worker hears that a user lost important work. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. An apology formula may be heard as institutional admission of causation. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Empathy about the event, with responsibility still unestablished. Use a form of sympathy that does not decide liability. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask who the target statement identifies as responsible. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 14: I apologize for the delay.

Source and situated reading. “I apologize for the delay.” A project manager acknowledges the team delivered later than promised. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Weakening it to generic sympathy erases responsibility for a missed commitment. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An explicit apology for timing, with accountability that should remain visible. Preserve both the admission of delay and the polite repair. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the speaker accepts that the delay occurred. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 15: I should have told you sooner.

Source and situated reading. “I should have told you sooner.” A friend recognizes a failure to communicate. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. An impersonal construction can erase ownership of the mistake. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. First-person retrospective responsibility and regret. Keep the speaker as agent without adding motives. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask who should have communicated. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 16: We regret that your application was unsuccessful.

Source and situated reading. “We regret that your application was unsuccessful.” An institution sends a standard decision notice. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A friendly paraphrase may imply there is still a pending decision. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A formula of regret accompanying a final negative outcome. Retain decision status and make no unmentioned appeal promises. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the application was accepted. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 17: I’m sorry to hear about your loss.

Source and situated reading. “I’m sorry to hear about your loss.” A person responds to news of a bereavement. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. An overly literal apology suggests culpability for the loss. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Compassion and acknowledgment, not responsibility. Prefer a genuine target condolence formula suitable for the relationship. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether a condolence or confession is communicated. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 18: Please accept my apology for interrupting.

Source and situated reading. “Please accept my apology for interrupting.” A panelist stops another speaker mid-explanation. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A casual ‘excuse me’ can be mistaken for a request to start speaking again. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A repair for a completed interruption, with explicit ownership. Keep the completed action and the attempt to restore the floor. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the speaker is repairing conduct or seeking permission. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Thanks, praise and social evaluation

This domain turns thanks, praise and social evaluation into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 19: Thank you for flagging that issue.

Source and situated reading. “Thank you for flagging that issue.” A reviewer appreciates a colleague’s defect report. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Literal flag imagery can obscure recognition of useful reporting. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Gratitude for timely warning, not praise for finding a fault in another person. Use the target convention for acknowledging a helpful report. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether reporting future defects feels welcome. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 20: That is an interesting approach.

Source and situated reading. “That is an interesting approach.” A senior reviewer says this after an unconvincing presentation. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Assuming praise can hide polite skepticism; assuming sarcasm can invent hostility. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A potentially ambiguous assessment whose tone and context matter. Keep ambiguity if evidence is thin, or clarify with the speaker. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask if the listener hears endorsement or a neutral acknowledgement. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 21: You’ve been very patient.

Source and situated reading. “You’ve been very patient.” A service employee responds after a long wait. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A literal characterization can sound condescending if the user had no choice. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Recognition of endurance and inconvenience. Use natural appreciation rather than a moral evaluation of personality. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask if the customer feels thanked or judged. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 22: I owe you one.

Source and situated reading. “I owe you one.” A teammate thanks someone for covering a shift. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Treating it as a legally enforceable debt invents obligation. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Informal gratitude with an offer of reciprocity. Use idiomatic relational gratitude, preserving the informal frame. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether a specific repayment is contractually due. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 23: Well done, everyone.

Source and situated reading. “Well done, everyone.” A coordinator closes a successful joint event. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Changing it to singular praise excludes some contributors. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Collective acknowledgement of group achievement. Maintain inclusive scope and do not invent individual rankings. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask who receives praise. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 24: That’s quite a choice.

Source and situated reading. “That’s quite a choice.” A companion comments on an unusual design without further evaluation. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Automatic praise or condemnation may overresolve an intentionally open remark. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An evaluative reaction whose polarity is context-dependent. Preserve the restrained evaluation unless intonation supplies stronger evidence. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask two listeners whether the judgment is clearly positive or negative. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Warnings, permissions and risk communication

This domain turns warnings, permissions and risk communication into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 25: You may enter after the light turns green.

Source and situated reading. “You may enter after the light turns green.” A device gate indicates when access is permitted. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Confusing ‘may’ with possibility can leave users unsure about authorization. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Permission conditional on a state change. Use target permission syntax and retain the gating condition. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask if entry before the green signal is allowed. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 26: Do not restart while the update is running.

Source and situated reading. “Do not restart while the update is running.” A screen warns about a device software update. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A polite suggestion version can invite unsafe experimentation. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A prohibition limited to the duration of the update. Keep the negative instruction and temporal condition. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether restart is ever allowed during processing. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 27: You might want to back up your draft.

Source and situated reading. “You might want to back up your draft.” A writing tool offers optional precautionary advice. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Making it compulsory invents a platform requirement. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Advisory action without a claim that backup already exists. Keep advice distinct from system-enforced duty. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the user is forced to back up. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 28: This area may be slippery.

Source and situated reading. “This area may be slippery.” A sign alerts visitors to a possible surface hazard. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Translating ‘may’ as certain misstates knowledge; removing the hazard makes it ineffective. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A possibility-based warning that informs prudent behavior. Preserve the uncertainty and the hazard, not a false guarantee. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether a hazard is possible and which behavior is prudent. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 29: Authorized staff only.

Source and situated reading. “Authorized staff only.” A marked door separates general and restricted access. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A loose translation as ‘staff preferred’ weakens access limits. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A role-based entry restriction. Keep the exclusion and avoid suggesting staff can bring unauthorized guests. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether an ordinary visitor may enter. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 30: Use this feature only if you understand the risks.

Source and situated reading. “Use this feature only if you understand the risks.” An advanced setting appears behind a caution notice. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A friendly simplification may erase the condition on use. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A conditional instruction connecting choice to informed understanding. Retain the condition and the fact that risks exist. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether mere curiosity is enough to proceed. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Evidence, probability and uncertainty

This domain turns evidence, probability and uncertainty into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 31: The figures appear to be consistent.

Source and situated reading. “The figures appear to be consistent.” An analyst has performed preliminary checks, not a complete audit. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. ‘Are consistent’ removes the observational limitation. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Tentative assessment from present evidence. Choose a target evidential marker that preserves tentative confidence. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the report claims definitive verification. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 32: The change may have caused the error.

Source and situated reading. “The change may have caused the error.” A development team investigates two simultaneous events. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Turning possibility into causation invents a root-cause finding. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A hypothesis awaiting tests. Keep the uncertainty and temporal reasoning separate. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether causation has been established. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 33: According to the witness, the door was open.

Source and situated reading. “According to the witness, the door was open.” A report records testimony without independent confirmation. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Removing attribution makes the report author endorse the claim. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Secondhand evidence explicitly sourced to a witness. Preserve reporting frame and avoid stronger certainty. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask who observed the door. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 34: This result is likely under those assumptions.

Source and situated reading. “This result is likely under those assumptions.” A technical note forecasts an outcome under modeled conditions. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Omitting the assumptions makes the prediction unconditional. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A probability judgment dependent on stated premises. Keep the likelihood and the conditional scope. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask what happens if assumptions change. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 35: We have not ruled out another explanation.

Source and situated reading. “We have not ruled out another explanation.” An investigation remains incomplete. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Paraphrasing as ‘another cause exists’ asserts more than the source. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Epistemic openness without endorsement of a specific alternative. Retain the negative finding about exclusion, not an invented cause. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether a second explanation is confirmed. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 36: As far as we know, no records were lost.

Source and situated reading. “As far as we know, no records were lost.” A team reports status after a partial check. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A definitive ‘none were lost’ erases the limits of available knowledge. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A bounded reassurance reflecting current evidence. Preserve the knowledge boundary and the negative statement. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the statement is absolute or provisional. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Consent, agreement and authorization

This domain turns consent, agreement and authorization into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 37: I understand the explanation.

Source and situated reading. “I understand the explanation.” A participant confirms comprehension after a lengthy briefing. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Replacing comprehension with consent creates a false authorization. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An acknowledgement of understanding, not necessarily approval. Keep understanding separate from agreement and subsequent action. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the person has actually agreed. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 38: I agree with the diagnosis, not the remedy.

Source and situated reading. “I agree with the diagnosis, not the remedy.” A team member accepts an analysis but challenges a proposal. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Collapsing the sentence to ‘I agree’ loses crucial scope. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Partial agreement with an explicit boundary. Retain the object of agreement and the rejected remedy. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask which part was accepted. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 39: Go ahead, provided the backup is complete.

Source and situated reading. “Go ahead, provided the backup is complete.” A technical lead authorizes a deployment conditionally. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Removing the condition turns provisional authorization into open permission. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Consent dependent on a prerequisite. Keep the prerequisite linked to the action. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether deployment before backup completion is authorized. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 40: I’ll think about it.

Source and situated reading. “I’ll think about it.” A person responds to a purchase proposal. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Calling this an acceptance invents a transaction. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Deferral of a decision, not consent. Use a target formula that leaves the decision pending. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the purchase can be processed. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 41: That works for me if everyone else agrees.

Source and situated reading. “That works for me if everyone else agrees.” A committee member expresses conditional support. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Removing ‘everyone else’ falsely records a group decision. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Individual acceptance conditional on broader agreement. Preserve both personal stance and collective-decision condition. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether collective approval is already obtained. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 42: Noted.

Source and situated reading. “Noted.” An attendee acknowledges a message in a group thread. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A generic ‘approved’ confuses receipt with endorsement. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Acknowledgement of information, without a decision. Use a neutral acknowledgement marker in the target channel. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the note authorizes the plan. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Clarification, repair and reference tracking

This domain turns clarification, repair and reference tracking into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 43: Sorry, which version did you mean?

Source and situated reading. “Sorry, which version did you mean?” A document editor encounters two drafts with similar labels. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Treating ‘sorry’ as culpability can distract from clarification. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A polite request to resolve reference ambiguity. Preserve the need to identify a version; do not guess. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask which file the recipient should open. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 44: I meant the earlier draft, not the final one.

Source and situated reading. “I meant the earlier draft, not the final one.” A speaker corrects a previous ambiguous reference. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Smoothing away the correction hides the possibility of action on the wrong file. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An explicit replacement of one referent by another. Keep correction polarity and version distinction visible. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask which version is now intended. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 45: When you say ‘complete’, do you mean submitted?

Source and situated reading. “When you say ‘complete’, do you mean submitted?” A coordinator checks whether an administrative milestone was reached. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Substituting the translator’s preferred definition for the participant’s obscures the question. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A metalinguistic clarification about task status. Keep the queried term distinct from the possible interpretation. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the coordinator knows submission occurred. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 46: Wait, that’s not what I was proposing.

Source and situated reading. “Wait, that’s not what I was proposing.” A negotiator interrupts a mistaken summary. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A calm paraphrase can make the disagreement disappear. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Immediate rejection of a misrepresentation. Preserve disagreement without adding accusations of bad faith. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the earlier summary still stands. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 47: Let me correct one detail.

Source and situated reading. “Let me correct one detail.” A presenter revises a number during a live explanation. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Removing the repair could leave incompatible versions untraceable. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A public correction that changes the authoritative next value. Preserve the correction signal and revised detail. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask which number should be used going forward. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 48: Are we talking about the same customer?

Source and situated reading. “Are we talking about the same customer?” Two teams appear to be using the same first name for different accounts. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A yes-no question about agreement may obscure reference identity. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A disambiguation request before proceeding with data. Retain the identity uncertainty and avoid revealing unrelated information. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the parties have verified the referent. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Humour, sarcasm and play

This domain turns humour, sarcasm and play into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 49: Well, that went perfectly.

Source and situated reading. “Well, that went perfectly.” A speaker watches a demonstration fail in a public rehearsal. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Literal praise reverses obvious irony. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A negative assessment conveyed through contrast with the event. Use target irony only if the audience can recover the failure context. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the target sounds critical despite positive words. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 50: Nice timing!

Source and situated reading. “Nice timing!” A friend arrives just after a group has left. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Without context, this could be sincere praise rather than teasing. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Possible ironic complaint about lateness. Preserve the playful or critical frame based on observable interaction. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask if the listener hears praise or reproach. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 51: You are a lifesaver.

Source and situated reading. “You are a lifesaver.” Someone thanks a colleague for fixing a small formatting error. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Literal life-saving language can imply a medical emergency. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Hyperbolic appreciation rather than factual rescue. Choose an idiomatic expression of strong gratitude proportionate to context. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask if anyone thinks a life was actually at risk. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 52: Great, another password reset.

Source and situated reading. “Great, another password reset.” A user expresses frustration after a repeated technical interruption. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Literal positive sentiment hides dissatisfaction. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Sarcastic disapproval of repeated inconvenience. Preserve the negative evaluation and repetition cue. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the user sounds pleased about the reset. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 53: The printer has decided to take a holiday.

Source and situated reading. “The printer has decided to take a holiday.” A team member jokes about a stalled office printer. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Word-for-word anthropomorphism can be awkward or falsely interpreted as policy. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Playful personification of equipment failure. Adapt the humorous failure image without inventing service status. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask if the actual problem, printer inactivity, remains clear. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 54: I see the spreadsheet is feeling adventurous.

Source and situated reading. “I see the spreadsheet is feeling adventurous.” A reviewer points to unexpected totals in an informal team chat. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. An earnest translation about emotion changes the referent. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Light teasing that draws attention to anomalous data. Keep the target’s attention on the anomaly and its low-stakes joking tone. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether a reader notices the suspicious figures. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Discourse markers, framing and rhetorical moves

This domain turns discourse markers, framing and rhetorical moves into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 55: Actually, the deadline was moved.

Source and situated reading. “Actually, the deadline was moved.” A speaker corrects an earlier schedule claim. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Removing ‘actually’ can understate that this is a correction. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A counter to prior information, not merely a new date. Use a target correction marker or explicit contrast. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the earlier deadline is still considered valid. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 56: Well, I’m not convinced.

Source and situated reading. “Well, I’m not convinced.” A participant responds cautiously to a persuasive presentation. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Treating ‘well’ as meaningless filler can erase hesitation. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A dispreferred disagreement with measured stance. Preserve reluctance and lack of conviction. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the listener hears support or resistance. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 57: Anyway, let’s return to the evidence.

Source and situated reading. “Anyway, let’s return to the evidence.” A chair ends a distracting discussion. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Word-by-word substitution misses the topic-shift function. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A discourse reset and procedural redirection. Use a natural transition marker and clear topic. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether participants move back to evidence. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 58: So, what does that leave us with?

Source and situated reading. “So, what does that leave us with?” A facilitator summarizes the consequences of several eliminated options. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A literal inventory question may miss the inference request. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A prompt to draw a conclusion from previous discussion. Retain the inferential link to what came before. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether respondents list options or produce a conclusion. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 59: On the other hand, the risks remain.

Source and situated reading. “On the other hand, the risks remain.” An author balances an earlier favorable claim. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Dropping contrast makes the text one-sided. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A countervailing consideration with retained risk. Preserve argumentative opposition, not just the risk statement. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the target reader sees a balanced assessment. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 60: In other words, this is a temporary fix.

Source and situated reading. “In other words, this is a temporary fix.” A speaker reformulates technical details for non-specialists. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Treating it as a new independent claim might suggest two separate fixes. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A reformulation of an earlier conclusion. Use an explanation marker and preserve temporary status. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the audience sees one solution or two. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Address, hierarchy and social distance

This domain turns address, hierarchy and social distance into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 61: Could you approve this, please?

Source and situated reading. “Could you approve this, please?” A junior employee writes to the person with formal sign-off authority. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Target address forms may wrongly invert hierarchy. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A request directed to the correct approving role. Confirm roles before choosing relationship-sensitive pronouns or honorifics. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask who has authorization to sign. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 62: Thanks, team.

Source and situated reading. “Thanks, team.” A project leader posts a closing note to several contributors. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Replacing the collective address with one honorific erases the group. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Informal collective recognition. Use a natural inclusive group address. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask who is being thanked. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 63: Dear colleague,

Source and situated reading. “Dear colleague,” A formal announcement is sent to peers across several departments. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Over-familiar salutations can imply a personal relationship. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Professional address without specific rank or gender inference. Choose a neutral professional target salutation. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the message feels private or institutional. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 64: Professor, may I ask a question?

Source and situated reading. “Professor, may I ask a question?” A student asks permission to interrupt a lecture. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Dropping the address may lose role information that is active in this genre. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A deference-sensitive request to speak. Use target forms appropriate to academic instruction without inventing extra titles. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask who is requesting the floor. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 65: Please address your response to the committee.

Source and situated reading. “Please address your response to the committee.” An official notice identifies the deciding body. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Rendering ‘committee’ as an individual introduces incorrect accountability. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A routing direction to a collective institutional recipient. Keep recipient identity and procedural role. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask to whom the written response should be sent. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 66: I’m speaking for myself, not for the organization.

Source and situated reading. “I’m speaking for myself, not for the organization.” A participant explicitly limits representation at a public event. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Switching to ‘we’ can imply official endorsement. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Personal viewpoint with institutional nonrepresentation. Preserve first-person scope and disclaimer. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the organization is committed by the statement. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Presupposition, loaded questions and implied claims

This domain turns presupposition, loaded questions and implied claims into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 67: When did the file disappear?

Source and situated reading. “When did the file disappear?” An investigator asks about a missing document, but its loss has not been verified. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A target answer might concede the file vanished even if it is merely misplaced. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A question carrying a presupposition about an event. If accuracy matters, clarify whether disappearance is established. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the target question assumes a proven loss. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 68: Have you stopped sharing the code?

Source and situated reading. “Have you stopped sharing the code?” A manager asks about a suspected practice. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A literal target can imply the practice definitely happened before. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A polar question whose wording presupposes a prior action. Preserve or flag the premise rather than asserting guilt. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the recipient can deny ever sharing it. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 69: Even Alex agreed.

Source and situated reading. “Even Alex agreed.” A speaker notes unexpected support. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Losing ‘even’ removes the expectation contrast and potential implication about Alex. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An additive assertion marked as surprising. Preserve the scalar focus without inventing a negative judgment. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask which agreement is presented as unexpected. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 70: Only the coordinator can approve it.

Source and situated reading. “Only the coordinator can approve it.” A policy clarifies limited approval rights. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Moving ‘only’ can exclude the wrong actor or action. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An exclusivity claim about authority. Scope the limiting expression to the coordinator. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether other staff can approve. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 71: We still need another review.

Source and situated reading. “We still need another review.” A team checks a document after two rounds. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Dropping ‘still’ may erase the continuity of an unmet requirement. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A continuing obligation despite previous work. Keep persistence and review count distinct. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the review requirement has already been fulfilled. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 72: You forgot the attachment again.

Source and situated reading. “You forgot the attachment again.” An irritated colleague points out a repeated omission. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Removing ‘again’ softens the accusation; adding stronger language invents intent. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A claim of repetition with interpersonal criticism. Retain repetition without implying deliberate neglect. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether more than one omission is alleged. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Offers, promises, plans and future actions

This domain turns offers, promises, plans and future actions into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 73: I can take the next shift.

Source and situated reading. “I can take the next shift.” A volunteer offers to cover a roster gap. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Translating ability alone can conceal a practical offer. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A voluntary offer of service. Keep the offer and identify if confirmation is still needed. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the roster can be assigned immediately. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 74: I’ll try to finish tonight.

Source and situated reading. “I’ll try to finish tonight.” A freelancer gives a cautious effort commitment. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Turning ‘try’ into ‘will finish’ creates a deadline guarantee. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Commitment to attempt, not certain completion. Preserve the effort/outcome distinction. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether delivery is guaranteed. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 75: We intend to review the proposal.

Source and situated reading. “We intend to review the proposal.” A committee outlines its future plan. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Replacing intention with an approved decision misstates status. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A planned review, not approval of the proposal. Keep future intent and separate review from approval. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask if the proposal has been accepted. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 76: I promise to send the revised version.

Source and situated reading. “I promise to send the revised version.” A writer personally guarantees a delivery action. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Weakening to ‘may send’ evades the promise. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An explicit personal commitment. Preserve speaker, action and degree of commitment. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether failure to send would breach an expressed promise. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 77: We hope to reopen soon.

Source and situated reading. “We hope to reopen soon.” A service notice follows temporary closure. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. ‘Will reopen tomorrow’ invents both date and certainty. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An aspiration and vague timeframe. Use target hope language and do not manufacture a date. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether a definite reopening time exists. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 78: Let me check and get back to you.

Source and situated reading. “Let me check and get back to you.” A support representative cannot yet verify a policy. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Rendering it as an immediate answer invents knowledge. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An interim commitment to investigate and respond. Keep the follow-up action and unresolved status. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask what the reader expects to happen next. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Deadlines, sequencing and priority

This domain turns deadlines, sequencing and priority into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 79: Please respond when convenient.

Source and situated reading. “Please respond when convenient.” A nonurgent request allows the recipient to choose timing. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Fixing a deadline can create unwarranted pressure. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A flexible request with no specified due date. Maintain open timing while preserving the request itself. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether tomorrow is required. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 80: We need the final text by noon.

Source and situated reading. “We need the final text by noon.” A production release depends on a locked file. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Softening ‘need’ into preference can miss a real scheduling dependency. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A time-bound necessity in the project. Retain deadline and identify source of authority. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask what happens if the text arrives afterward. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 81: Check the seal before opening.

Source and situated reading. “Check the seal before opening.” A procedure has an order-sensitive verification step. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Reordering clauses can invite opening before inspection. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A sequence constraint. Use a construction where verification precedes access. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask which action occurs first. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 82: This is a priority, not an emergency.

Source and situated reading. “This is a priority, not an emergency.” A leader wants prompt work without escalating incident protocols. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Translating ‘priority’ as ’emergency’ can trigger incorrect procedures. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. An elevated but bounded urgency level. Preserve the distinction between importance and immediate crisis. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether emergency response is required. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 83: We can wait until the next cycle.

Source and situated reading. “We can wait until the next cycle.” A team decides not to use an exceptional release window. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. An overly eager translation can imply imminent launch. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Permission to defer action. Keep the future window and nonurgent choice. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask if work must launch now. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 84: Please do not send anything until I confirm.

Source and situated reading. “Please do not send anything until I confirm.” A sender asks others to hold outbound messages. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Deleting the confirmation condition can turn the hold into an indefinite ban. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A temporary prohibition dependent on later authorization. Maintain the hold and the confirmation event. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether a message may be sent before explicit confirmation. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Disagreement, negotiation and delicate corrections

This domain turns disagreement, negotiation and delicate corrections into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 85: I hear your concern, but the data do not support it.

Source and situated reading. “I hear your concern, but the data do not support it.” A review chair acknowledges someone’s worry without endorsing the claim. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. ‘I agree’ falsely turns empathy into substantive agreement. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Respectful acknowledgement plus evidence-based disagreement. Keep the recognition and the evidence judgment separate. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the chair accepts the factual claim. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 86: Could we revisit the fee?

Source and situated reading. “Could we revisit the fee?” A client opens a negotiation without demanding a discount. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A price-cut instruction removes mutual choice. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A request to discuss an agreed or proposed amount. Preserve open negotiation and avoid inventing the result. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the fee has changed. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 87: That is not the trade-off we agreed to.

Source and situated reading. “That is not the trade-off we agreed to.” A partner disputes a summary of a prior decision. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A literal technical explanation can lose the objection. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A challenge to an alleged agreement. Keep the claim of disagreement without declaring fraud. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the past agreement is in dispute. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 88: I can agree to the scope, but not the schedule.

Source and situated reading. “I can agree to the scope, but not the schedule.” A vendor negotiates two separate terms. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A short ‘agreed’ can create false full acceptance. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Partial agreement divided by condition. Maintain distinct acceptance and rejection domains. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask which contractual terms remain open. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 89: I understand your point; here’s another possibility.

Source and situated reading. “I understand your point; here’s another possibility.” A participant offers a nonconfrontational alternative. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Treating it as praise for the original proposal misses the new option. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Acknowledgement followed by a proposal. Keep the sequencing and optional character. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether a second option is now under discussion. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 90: We should not call this a failure yet.

Source and situated reading. “We should not call this a failure yet.” An evaluator cautions against a premature conclusion. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Rendering as ‘it succeeded’ flips uncertainty into optimism. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A restraint on premature negative classification. Preserve the epistemic caution without inventing success. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether the outcome has been determined. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Diagnostic domain: Interfaces, automated messages and short replies

This domain turns interfaces, automated messages and short replies into a practical translation problem. Read the source line, identify the turn it occupies, then decide what a recipient is authorized or expected to conclude. A target sentence is not validated simply because its grammar and terminology are fluent. For this domain, test how the same action would be understood when the participants’ relationship, medium or stakes change. Keep the speaker’s actual level of obligation, certainty and interpersonal stance in view.

Case 91: Request received.

Source and situated reading. “Request received.” A portal confirms receipt of an uploaded application. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. ‘Approved’ would turn intake status into an outcome. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Acknowledgement of receipt only. Use a target status label distinct from review or approval. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask if a decision has been made. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 92: Continue without saving?

Source and situated reading. “Continue without saving?” An interface presents a potentially destructive choice. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Softening the question can hide data-loss consequences. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A confirmation prompt for a risky action. Keep the action and unsaved state visible. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask what would be lost by continuing. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 93: You’re all set.

Source and situated reading. “You’re all set.” A tool displays this only after account setup, not after payment verification. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. Interpreting it as ‘payment confirmed’ invents another completed process. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A context-limited completion message. Specify the completed task only when the interface state verifies it. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask what exactly is complete. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 94: Maybe later.

Source and situated reading. “Maybe later.” A person selects a notification-permission option. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. An overly forceful negative could represent a permanent opt-out. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A deferral with the option of future reconsideration. Preserve reversible postponement if the product supports it. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether consent has been permanently denied. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 95: Action required.

Source and situated reading. “Action required.” A security notification contains a mandatory identity check. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A casual FYI message weakens the need for action. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. A directive to complete a specific step. Keep requirement and link to accurate next action. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether ignoring the notice is acceptable. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Case 96: Thanks!

Source and situated reading. “Thanks!” A teammate sends a one-word acknowledgement after receiving a shared link. Begin by identifying the explicit proposition, its place in the dialogue and the practical response it makes relevant. Do not begin by selecting words in the target language; first establish which part is factual content and which part is interpersonal action. A statement with familiar vocabulary can still perform a social act that a purely literal translation fails to represent.

Hidden error. A translator might read this as acceptance of a proposed contract. The mistake is consequential because an audience does not receive the sentence as an abstract grammatical specimen. It interprets a relationship, a claim about what is known, and an invitation or restriction on what can happen next. Compare the defective interpretation with the original situation. Is a new duty being created, an existing duty removed, a source of evidence lost, or a person’s willingness misrepresented?

Function to carry across languages. Simple appreciation without necessarily agreeing to any terms. Choose a brief social acknowledgement, not a legal approval formula. Preserve the difference between what the utterance asserts and what the listener is reasonably meant to infer. Translate into a construction the target audience actually uses for this act. If two viable constructions give different levels of force, record why one fits the speaker, the listener and the medium more closely than the other.

Reader diagnostic. Ask whether any proposal was accepted. Also ask which preceding turn would make the interpretation change. A defensible translation should survive that counterfactual test where the source itself is stable; if it does not, the reviewer should explain the missing contextual cue. Keep this example in a contrast set with a direct imperative, a softer suggestion and a neutral factual statement. These contrasts reveal pragmatic drift that simple back-translation often conceals.

Transfer: from single sentences to paragraphs, products and conversations

Translation quality must be evaluated at the scale at which interaction actually occurs. In a conversation, an indirect refusal can be confirmed by the next speaker’s acceptance of the boundary; in a one-line interface, no later turn exists to rescue an ambiguity. A translator therefore needs more explicit support when the target line is detached from its original dialogue. The same question can be harmless in a story, legally relevant in a consent form and operationally decisive in a deployment checklist.

A usable transfer procedure starts with a unit of action rather than a unit of text. For every important utterance, mark its speaker, recipient, context, proposition, action, uncertainty, evidence, consequence and required response. At paragraph scale, trace how each claim changes the state of the discussion. At document scale, track who has authority to commit an organization. At product scale, trace whether the words match the action the interface really performs.

For machine translation and AI-assisted translation, pass the system enough context to distinguish roles and discourse acts, but do not assume that adding context automatically guarantees a better result. In an October 2026 WMT paper, Huang and Tsai evaluated workplace requests with and without sender-recipient profiles and found no statistically detectable overall improvement in appropriateness in their small evaluated set; a role-reversal test nevertheless exposed relationship sensitivity that simple positive ratings often concealed. The practical lesson is to test the output and the evaluator’s sensitivity, not to treat either a prompt or fluent wording as self-validating.

The translator should keep the source as the evidentiary authority. A target-language cultural convention can justify restructuring a phrase or choosing another form of address, but it cannot justify inventing a promise, consent, verified fact or authoritative decision. When the underlying interaction is ambiguous, an explicit reviewer question may be more honest than a polished answer. Where a source line has legal, medical, safety or financial consequences, seek qualified domain review rather than trusting a generic stylistic judgement.

Practice exercises and a self-audit you can repeat

Workshop 1: One turn, three acts

Exercise. Choose a line that looks like a question. Place it after a request for information, after a deadline reminder and after a service complaint. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Explain how the function changes without changing the words, then identify which target expressions preserve each function. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 2: The relationship reversal

Exercise. Translate a polite request with the speaker and recipient as equals, then reverse a confirmed supervisory relationship. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Record which forms must change, which facts stay unchanged and which assumptions cannot be justified. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 3: The uncertain apology

Exercise. Compare empathy for an event, regret about a failure and an explicit admission of responsibility. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Write three target candidates and ask whether each assigns blame to the same actor. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 4: A refusal with an exit route

Exercise. Translate a clear refusal that also offers an alternative time, then a vague deferral with no date. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Check whether the target invents a fixed appointment or makes a refusal sound negotiable. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 5: Evidence versus confidence

Exercise. Translate a finding labeled reported, likely, possible and confirmed. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Ask an independent reader to rank the certainty; mismatched ranks reveal a stance defect. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 6: Consent versus receipt

Exercise. Place ‘noted’, ‘understood’, ‘agreed’ and ‘approved’ at the end of a workflow. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Predict what systems and people should be permitted to do after each response. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 7: Polarity under pressure

Exercise. Translate a negative question in a context where yes/no answer patterns differ. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Have a reviewer respond naturally, then verify that the resulting action matches the source. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 8: Discourse-marker recovery

Exercise. Remove ‘actually’, ‘well’, ‘still’ and ‘only’ from four sentences. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. State exactly which contrast, reluctance, persistence or scope the deletions lose. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 9: The live-dialogue test

Exercise. Translate six consecutive conversational turns instead of isolated sentences. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Track repair, interruption and response expectations; note any apparent agreement created by smoothing. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 10: The interface consequences test

Exercise. Translate a confirmation, a warning, a completed-status message and a postponement button. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Put each in its actual interface location and ask users what they believe has happened. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 11: The quotation boundary

Exercise. Translate a paragraph containing a disputed allegation attributed to a named witness. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Have a target reader identify the claimant, reporter and evidence status. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 12: The irony check

Exercise. Translate a sentence of praise in a context of failure and another context of genuine success. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Do not force a single ironic reading when the evidence changes. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 13: The audience channel

Exercise. Render the same update as an informal message and an official policy note. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Preserve the same facts but alter only socially appropriate presentation. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 14: The unexplained silence

Exercise. Write a report of a meeting in which a proposal receives no reply. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Avoid describing the silence as approval or rejection unless a verified procedural rule applies. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 15: The source-ambiguity gate

Exercise. Use an utterance with two plausible intentions and withhold the preceding turn. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Write the exact question that would resolve the uncertainty instead of guessing. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Workshop 16: The independent reception check

Exercise. Give a target-only reviewer a translation without the source and ask what action follows. Before drafting, record a hypothesis about the function that must survive. Identify which details are anchored in the source and which would require external confirmation. This creates a baseline against which the target can be assessed rather than relying on memory after the translation is complete.

Method. Compare that response with the source purpose and decide whether the divergence is a defect. Use two target alternatives that are each grammatical but differ in the social action they invite. Explain the contrast to a second reviewer without telling that reviewer which version you prefer. Ask for a specific prediction about how an ordinary recipient would respond; a preference for elegance alone is not enough.

Pass criterion. The selected target preserves the proposition, appropriate obligation or permission, attributed certainty, relationship signal and any deliberate ambiguity. A failed answer is not evidence that the reader is incompetent. Revisit the source interpretation, the assumptions in the translation brief and the concrete target-language construction, then repeat the test on the corrected version.

Research notes and authoritative reading

The research literature supports separating communicative function from literal wording. The Routledge Handbook of Translation and Pragmatics, edited by Rebecca Tipton and Louisa Desilla, maps pragmatic questions across translation and interpreting. Its value here is conceptual: the reader should notice that interlocutor relations and inferred meaning are legitimate objects of translation analysis, rather than treating them as optional decoration.

In the Proceedings of the Eleventh Conference on Machine Translation (October 2026), Huang and Tsai examined pragmatic appropriateness in workplace translation and showed how role-reversal and comparative judgments may reveal problems hidden by coarse positive ratings. This is a small study, not a universal claim that a particular AI system or profile strategy solves pragmatics. The guide treats it as a methodological caution about how to evaluate relationship-sensitive output.

Liu and Li’s 2026 study on interpreting lexical bundles describes stance, discourse-organizing and referential functions, with observed correspondences, additions and shifts. It illustrates why the organization of discourse deserves its own audit. Separately, Taguchi and Dennis investigate acceptances and refusals in prosody, warning against assuming that all the communicative work is visible in the transcript. A 2025 comparative study of requests, refusals and apologies reinforces the importance of interactional context; its language-pair findings should not be turned into stereotypes about every member of a language community.

The W3C Internationalization Tag Set 2.0 illustrates the technical need to mark translatability, terminology and localization-related metadata in multilingual content pipelines. It does not prescribe a universal politeness formula. Here it serves as a reminder to carry context and constraints through the workflow instead of allowing sentence-by-sentence translation to discard them.

• Huang and Tsai, Evaluating Pragmatic Appropriateness in English–Vietnamese Machine Translation with Dyadic Profiles (WMT, October 2026).
• Tipton and Desilla (eds.), The Routledge Handbook of Translation and Pragmatics (2019).
• Liu and Li, Interpreting lexical bundles: discourse function and correspondence patterns (2026).
• Taguchi and Dennis, Prosody of Preferred and Dispreferred Speech Acts, online publication July 26, 2023; journal volume 2025.
• Alindra and colleagues, Speech Acts Across Cultures: A Comparative Study of Request, Refusal, and Apology (December 14, 2025).
• W3C, Internationalization Tag Set 2.0 (2013).

Next routes: keep each learning owner distinct

This chapter belongs to Master Art of Translation | The Complete System for Moving Meaning Between Languages. That root holds the overall map for context, equivalence, terminology, revision, technology and quality assurance. Its role is to help readers find the correct next part of the translation system. This chapter’s narrower responsibility is to preserve speech acts and interpersonal consequences when language changes. The chapter immediately before it, Multilingual Comprehension Testing — Cognitive Interviews, Teach-Back and Reader Tasks, explains how to establish reader understanding systematically; use it when you need empirical reception evidence for the target.

Readers who cannot yet recognize or retrieve the words needed to express an idea should use the protected Vocabulary Learning Hub. That hub owns vocabulary acquisition, word meanings, examples and retrieval. Readers who need to understand English clause structure, discourse, register or grammar inside English should use How English Works V1.1. Neither protected owner needs to be rewritten to explain this crossing between languages. A source sentence can contain a perfectly familiar word while its function shifts across languages; that crossing is the reason this page exists.

The next translation task should become easier to name: identify the source proposition, determine the action it performs, recover the relationship and evidence that make the action intelligible, choose a fitting target expression, and verify the recipient’s predicted response. When a person does something different after reading the translation, investigate what moved. The central principle remains simple: change the language without silently changing permission, pressure, responsibility, uncertainty, dignity or consent.