Questions in translation do more than place a question mark at the end of a sentence. They define what information is missing, what alternatives are open, what answer is expected and whether the speaker is neutral, surprised, sceptical or seeking confirmation. “Did Lina call?” asks about truth of a proposition. “Who called?” asks for an identity. “Did Lina call or write?” presents alternatives. “You called, didn’t you?” often seeks confirmation while revealing a prior expectation.
People searching for how to translate yes/no questions, wh-questions, alternative questions, tag questions, embedded questions or question particles are solving an interrogative-structure problem. Languages form questions through inversion, particles, word order, intonation, morphology or special pronouns. A literal source order can be ungrammatical or can ask a different question in the target. Precision begins by identifying the question type and requested information.
This article owns the interrogative precision layer. It covers polar questions, wh-questions, alternative questions, tag questions, negative questions, echo questions, rhetorical questions, embedded interrogatives, question scope, presuppositions, answer systems and QA. It routes broader sentence structure to the existing owner and speech-act effects to the Speech Acts lane. The central rule is: identify exactly what the source asks the listener to supply or confirm before choosing target question form.
Yes/No Questions
The key is to identify the source relation before choosing target syntax. Polar questions ask whether a proposition is true, usually inviting confirmation or rejection. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is turning them into content questions or statements with rising tone. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to write proposition P and question ‘P true?’ For example, “Did Lina call?” asks whether Lina called. Before release, target should permit a yes/no-type response. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Wh-Questions
A precise translation treats this as a semantic-and-grammatical mapping problem. Wh-questions ask for a missing participant, time, place, reason, manner or quantity. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is moving the interrogative scope to the wrong constituent. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to replace wh-word with a blank slot For example, “Who called Lina?” asks for caller; “Who did Lina call?” asks for recipient. Before release, fill target blank with a sample answer. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Subject vs Object Wh-Questions
This becomes difficult when the target language packages the same meaning through a different construction. Some languages mark subject/object interrogatives differently or not at all. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is swapping actor and affected participant. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to identify grammatical and semantic role of wh-element For example, who called? vs who did she call? Before release, answer with named participants and check event. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Where Questions
The source wording is only the surface; the underlying event or clause relation is the invariant. Where can ask location, source, goal or route depending on motion predicate. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is using one target locative interrogative where language distinguishes from/to/at. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to label AT / FROM / TO / THROUGH For example, where are you? vs where are you going from/to? Before release, answer with location and relation. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
When Questions
A fluent target can still be wrong if this relation changes during restructuring. When can request clock time, date, duration boundary or event position. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is choosing target form that asks how long rather than when. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to identify answer type For example, “When did it start?” expects a time point, not duration. Before release, supply sample answer. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Why Questions
The key is to identify the source relation before choosing target syntax. Why seeks reason, purpose or cause depending on predicate and context. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is translating reason as purpose or vice versa. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to test because vs in order to answers For example, “Why did she leave?” may ask cause/reason; “Why are you here?” can ask purpose. Before release, target answer relation. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
How Questions
A precise translation treats this as a semantic-and-grammatical mapping problem. How can request manner, method, condition, degree or state. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is using generic target interrogative and losing dimension. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to predict expected answer For example, “How did you solve it?” asks method; “How tall?” asks degree. Before release, sample answer fits target. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
How Many vs How Much
This becomes difficult when the target language packages the same meaning through a different construction. Quantity questions interact with count/mass distinctions. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is using wrong target classifier or quantity form. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to resolve noun quantity model first For example, how many books vs how much water. Before release, answer with unit. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Alternative Questions
The source wording is only the surface; the underlying event or clause relation is the invariant. Questions such as ‘tea or coffee?’ ask the listener to choose among named alternatives. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is translating as yes/no question about the whole disjunction. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to label alternatives and expected answer set For example, “Did Lina call or write?” may seek which action. Before release, target should invite ‘called’ or ‘wrote’, not merely yes. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Inclusive-Or Ambiguity
A fluent target can still be wrong if this relation changes during restructuring. Some ‘or’ questions can be ambiguous between polar and alternative readings. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is forcing one interpretation without context. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to use intonation/context or preserve ambiguity For example, “Do you want tea or coffee?” usually alternative; “Can you read or write?” can ask whether either ability exists. Before release, identify natural answer type. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Tag Questions
The key is to identify the source relation before choosing target syntax. Tags seek confirmation while expressing expectations or stance. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is translating as neutral yes/no question. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to identify base assertion and expected polarity For example, “You sent it, didn’t you?” suggests speaker expects yes. Before release, target preserves confirmation-seeking force. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Negative Questions
A precise translation treats this as a semantic-and-grammatical mapping problem. Negative interrogatives can express surprise, expectation, challenge or persuasion. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is treating them as neutral information requests. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to identify expected answer and stance For example, “Haven’t you finished?” often expects that finishing should have happened. Before release, compare pragmatic pressure. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Echo Questions
This becomes difficult when the target language packages the same meaning through a different construction. Echo questions repeat or partially repeat prior speech to request clarification or express surprise. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is normalizing them into standard questions and losing interaction. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to identify repeated material and focus For example, “You bought what?” can express surprise/clarification. Before release, target should preserve echo function. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Rhetorical Questions
The source wording is only the surface; the underlying event or clause relation is the invariant. A question can make a claim rather than seek information. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is translating it as genuine open question. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to ask whether an answer is expected For example, “Who could have known?” often implies nobody could. Before release, preserve rhetorical reading. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Embedded Questions
A fluent target can still be wrong if this relation changes during restructuring. Interrogative content can be embedded under verbs such as know, ask, wonder and find out. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is using main-question word order inside a target embedded clause. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to separate interrogative meaning from matrix clause grammar For example, “I wonder where she is” differs from “Where is she?” Before release, target embedded syntax and scope. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Whether vs If
The key is to identify the source relation before choosing target syntax. English whether and if overlap in embedded polar questions but differ in some constructions. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is using a target conditional marker instead of interrogative complement. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to paraphrase as ‘the answer to the question P?’ For example, “I asked whether he came” reports a question, not a condition. Before release, matrix verb should take interrogative content. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Question Particles
A precise translation treats this as a semantic-and-grammatical mapping problem. Some languages mark polar questions with particles rather than inversion. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is copying English inversion or omitting required target marker. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to use target interrogative system naturally For example, a clause may keep declarative order plus final particle. Before release, native question pattern. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Question Scope
This becomes difficult when the target language packages the same meaning through a different construction. Only part of a sentence may fall under interrogation. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is asking about the wrong participant, time or embedded clause. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to mark the unknown slot For example, “Did Lina say Omar left?” can question whether she said it, not whether Omar left. Before release, state expected answer. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Presuppositions in Questions
The source wording is only the surface; the underlying event or clause relation is the invariant. Questions can take background propositions for granted. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is neutralizing or strengthening source assumptions. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to separate requested information from presupposed content For example, “When did you stop using it?” presupposes prior use and stopping. Before release, compare background status. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Answer Systems Across Languages
A fluent target can still be wrong if this relation changes during restructuring. Languages differ in how yes/no responses align with positive/negative questions. In practice, this means separating the source into its functional parts before drafting. The target may use a different complementizer, non-finite form, aspectual verb, path expression, question particle or clause order. That is acceptable only when the same participant roles, event boundaries, discourse assumptions and task demands remain recoverable to the target reader.
The common failure is translating isolated yes/no answers without the question context. This can survive ordinary proofreading because every individual target word may be grammatical and plausible. The error appears only when the reviewer asks what event completed, who controls the embedded action, what path the mover follows, or what information the question actually requests. In technical, educational, legal, research and professional contexts, such a shift can alter instructions, claims, chronology, responsibility or required responses.
A practical method is to translate question-answer pair together For example, responses to “Didn’t you go?” can map differently across languages. Before release, reconstruct proposition confirmed/denied rather than word yes/no. This converts intuition into a repeatable check. When the source leaves something ambiguous or underspecified, preserve or flag that uncertainty instead of forcing a neat target interpretation. The goal is not to keep the source grammar visible; it is to keep the source relation intact.
Worked Examples
Worked Example 1: Who Called vs Who Did She Call?
These two questions contain the same interrogative word but ask for different semantic roles. In the first, the unknown is the agent; in the second, the unknown is the affected/recipient participant.
A target language with little word-order change may rely on case or particles. The role, not English inversion, is the invariant.
Worked Example 2: Tea or Coffee?
In a café, Tea or coffee? normally asks the customer to choose an option, not to answer yes or no.
A target yes/no construction that means “Do you want tea-or-coffee?” may invite the wrong response. Preserve the alternative-question function.
Worked Example 3: I Wonder Where She Is
English embedded questions do not use ordinary main-clause inversion: where she is, not where is she. Other languages have their own embedded interrogative syntax.
The target must remain a complement of wonder rather than becoming a separate direct question.
Worked Example 4: Didn’t You Send It?
A negative question can seek confirmation while revealing an expectation that the item was sent.
If the target language’s yes/no system behaves differently under negation, translate the whole question-answer logic rather than mapping yes and no mechanically.
Practical Diagnostic Workflow
Classify each question as polar, wh/content, alternative, tag, echo, rhetorical or embedded. Then identify the unknown slot and the expected answer type: truth value, person, place, time, reason, manner, amount or choice. Mark any presupposition or stance that accompanies the question.
After translation, answer the target question yourself with a sample response. If the sample answer differs in type from what the source requests, the question structure has drifted. For embedded questions, also check that the interrogative remains embedded under the correct matrix predicate.
Recovery When Meaning Has Drifted
When a question has shifted, remove surface punctuation and reconstruct the requested information. Restore the unknown slot or alternative set first, then rebuild target interrogative syntax. This avoids fixing word order while leaving the wrong question scope.
Then check pragmatic force. A structurally correct negative or tag question can still be wrong if the target loses the speaker’s expectation, surprise or challenge. Route those cases through the speech-act and presupposition layers as needed.
Quality-Assurance Checklist
- Classify question type.
- Identify unknown slot.
- Preserve subject/object role.
- Distinguish location/source/goal questions.
- Check reason vs purpose.
- Resolve alternative vs polar or.
- Preserve tag-question expectation.
- Preserve negative-question stance.
- Keep echo/rhetorical function.
- Use correct embedded-question syntax.
- Separate whether from conditional if.
- Translate yes/no answers with question context.
Frequently Asked Questions
Should questions keep the same word order?
No. Languages form interrogatives differently. Preserve the requested information and question type, not English inversion.
What is an embedded question?
Interrogative content functioning inside a larger clause, as in ‘I wonder where she is’.
Why are alternative questions tricky?
Because ‘or’ can form a choice question or appear inside a yes/no question. Context and expected answer distinguish them.
Are negative questions neutral?
Often not. They can express expectation, surprise or pressure, which may need target pragmatic marking.
Can yes/no be translated without the question?
Sometimes, but negative questions are especially risky because languages align affirmative/negative responses differently.
Can AI change what is being asked?
Yes, especially when simplifying complex or embedded questions. Always test the target with a sample answer.
Where This Fits in the eduKate Translation Architecture
Use the existing sentence-structure owner for broad interrogative word order, Speech Acts and Communicative Force for requests/rhetorical uses, and Presupposition/Implicature for loaded questions. Count/mass and motion owners support quantity and spatial interrogatives.
Final Principle
Translate the information gap. Identify what the source asks the listener to supply or confirm, then choose the target interrogative form that asks for exactly that. Punctuation and word order can change; the requested answer must not.