An embedded question is a question-shaped clause used as part of a larger sentence: Ban san ko wa ya ɗauki makullin ba “I do not know who took the key.” The identity of the taker is still open, but the whole sentence is a statement about the speaker’s knowledge. Hausa normally marks that local interrogative domain with ko, places the wh-expression at the embedded left edge, and uses the same relative TAM series found in matrix wh-fronting, focus, and relative clauses. The basic contrast with clause-level fronting is developed in In-Situ versus Fronted Focus.
All unlabelled examples are (neutral Nigerian Standard Hausa). Ordinary boko does not mark lexical tone or vowel length. Question punctuation belongs to the entire sentence: an embedded question inside a declarative does not by itself force a final question mark.
Ko introduces the embedded question domain
With predicates of knowing, finding out, asking, remembering, or explaining, ko + wh-expression introduces the unresolved content.
Ban san ko wa ya ɗauki makullin ba.
I do not know who took the key.
Muna son mu san ko me kuka kawo.
We want to know what you all brought.
Ta tambaya ko a ina suka ajiye motar.
She asked where they had parked the car.
Ba su bayyana ko yaushe za su dawo ba.
They did not explain when they would return.
Here ko is an embedded-interrogative complementizer, roughly “whether/which answer.” It is not the separate coordinator ko “or,” and it is not part of a lexicalized indefinite such as kowa “everyone/anyone” or ko’ina “everywhere/anywhere.” The spaces matter: ko wa means “which person, in an embedded question,” while kowa is one word with a different distribution.
Relative TAM remains local to the embedded wh-dependency
Fronting a wh-expression inside the embedded clause triggers relative completive or relative continuative marking there. It does not change the TAM of the matrix clause.
| Embedded function | Example | Diagnostic |
|---|---|---|
| subject, completed | Ban san ko su wa suka iso ba. | suka, not sun |
| object, completed | Ban san ko me yaran suka saya ba. | suka, not sun |
| object, ongoing | Ta tambaya ko me suke gyarawa. | suke, not suna |
| future | Ba mu san ko wa zai zo ba. | future has no separate relative series |
Ban san ko su wa suka iso da wuri ba.
I do not know who all arrived early.
Ta tambaya ko me suke gyarawa a ɗakin baya.
She asked what they were repairing in the back room.
The matrix negatives ban ... ba and ba mu ... ba remain ordinary matrix grammar. Relative suka/suke belongs only to the clause whose fronted wh-expression it licenses. Singular ordinary boko can hide the distinction because relative and plain forms may both be written ya or ta; plural examples expose it.
The same family resemblance links three constructions:
| Construction | Example |
|---|---|
| embedded wh-question | ko me suka saya |
| fronted focus | Littattafai ne suka saya. |
| relative clause | littattafan da suka saya |
The three are not interchangeable in meaning, but each connects a prominent expression or operator to a position interpreted inside a relative-TAM clause.
Local embedded scope versus matrix scope
Scope tells you which question the wh-expression answers. In a declarative such as Ban san ko wa ya zo ba, wa has scope only inside the complement of san: the speaker asserts ignorance and leaves the visitor’s identity unresolved.
A matrix polar question can contain the same embedded question:
Ka san ko wa yake kula da ɗakin?
Do you know who looks after the room?
Strictly, the whole question asks whether the addressee knows; eh or a’a can answer its matrix yes/no force, although cooperative speakers often supply the person’s identity too. Compare a matrix content question whose wh-expression asks directly for a person associated with the embedded proposition:
Wa ka ce ya ɗauki makullin?
Who did you say took the key?
Here initial wa takes scope over the whole utterance: the requested answer is a person, not merely “yes, I said it.” This is a long dependency and should not be treated as permission to pull a wh-expression freely out of every subordinate clause. In particular, an embedded ko question is already its own interrogative domain. For reliable production, keep its wh-expression locally after ko rather than extracting it across that boundary.
Wh-questions and embedded focused assertions are different
An embedded question leaves a value open. An embedded focused assertion identifies a value and can follow cewa “that.”
Ban san ko wa ya turo saƙon ba.
I do not know who sent the message. (open embedded question)
Na tabbatar da cewa Amina ce ta turo saƙon.
I confirmed that Amina was the one who sent the message. (embedded focused assertion)
In the first sentence, wa does not normally need ne as an interrogative auxiliary; the wh-expression and relative TAM establish the dependency. In the second, Amina ce identifies a feminine singular value, and ta supplies subject/TAM marking. Ce is not an agreeing verb and carries no past tense; past interpretation comes from the clause and context.
The focused assertion generally backgrounds the proposition that someone sent the message and settles or corrects the identity. The embedded question does not settle it.
Neutral statement, embedded questions, and correction
Begin with Yaran sun kawo rahoto ofis “The children brought a report to the office.”
| Function | Hausa |
|---|---|
| neutral statement | Yaran sun kawo rahoto ofis. |
| embedded subject question | Ban san ko su wa suka kawo rahoton ofis ba. |
| embedded object question | Ban san ko me yaran suka kawo ofis ba. |
| matrix question about knowledge | Ka san ko me yaran suka kawo? |
| embedded focused assertion | Na ji cewa rahoto ne yaran suka kawo. |
| clefted correction | Ba hotuna ba ne yaran suka kawo; rahoto ne. |
The first row reports an event with plain sun. The embedded questions use ko and relative suka. The last row no longer asks: it corrects one proposed object while retaining the bringing event as background.
A natural dialogue
—Ka san ko wa ya kawo wannan fayil ɗin? —Ban sani ba. Zan tambayi mai gadi ko wa ya shigo da shi. —Musa ne ya kawo shi? —A’a, na tabbatar da cewa Amina ce ta kawo shi. —To, ka san ko me take jira? —Tana jiran a sa wa fayil ɗin hannu.
—Do you know who brought this file? —I don’t know. I’ll ask the guard who came in with it. —Was it Musa who brought it? —No, I confirmed that Amina brought it. —Then do you know what she is waiting for? —She is waiting for the file to be signed.
The first and second ko wa clauses preserve an unresolved embedded identity. The focus question proposes Musa, and the cewa Amina ce clause supplies a correction. The final ko me opens a new embedded question.
Common Mistakes
1. Omitting ko in the productive indirect-question pattern
❌ Ban san wa ya ɗauki makullin ba.
Noncanonical for the productive indirect-question pattern taught here; the embedded question boundary is not marked with ko.
✅ Ban san ko wa ya buɗe ƙofar ba.
I do not know who opened the door.
2. Keeping plain plural TAM after embedded wh-fronting
❌ Ban san ko su wa sun zo ba.
Incorrect canonical embedded subject question; sun is plain completive.
✅ Ban san ko su wa suka zo ba.
I do not know who all came.
3. Letting ne replace subject–TAM marking
❌ Ban san ko wa ne kira ba.
Incorrect for ‘I do not know who called’; ne cannot replace the subject–TAM form ya.
✅ Ban san ko wa ya kira ba.
I do not know who called.
4. Giving an embedded question automatic sentence-final question force
❌ Ban san ko me suka saya ba?
Wrong punctuation for the intended declarative ‘I do not know...’; the question mark changes the force of the whole utterance.
✅ Ban san ko me suka saya ba.
I do not know what they bought.
5. Confusing ko wa with lexicalized kowa
❌ Ban san kowa ya zo ba.
Incorrect for ‘I do not know who came’; kowa is the lexicalized indefinite/universal ‘everyone/anyone.’
✅ Ban san ko wa ya zo ba.
I do not know who came.
Key Takeaways
- Embedded content questions normally begin with ko + wh-expression inside the complement selected by knowing, asking, discovering, and similar predicates.
- The embedded fronted wh-expression selects relative TAM in its own clause: suka/suke, not plain sun/suna.
- A declarative containing an embedded question ends with a period; only matrix question force licenses a final question mark.
- Local ko wa... scope differs from matrix Wa ka ce...? scope, and an embedded ko domain is not a free extraction route.
- Ne/ce identify values in focused assertions; they are not interrogative auxiliaries and do not replace subject–TAM marking.
Now practice Hausa
Reading grammar gets you part of the way. The exercises are where it sticks — free, no signup needed.
Start learning Hausa →Related Topics
- Multiple Wh-ExpressionsA2 — Ask about two unknowns in one Hausa question by fronting one wh-expression and leaving the other in its natural clause position.
- In-Situ versus Fronted FocusB1 — Choose between leaving new information in its ordinary position and fronting a Hausa constituent with relative TAM for stronger identification or correction.
- Focus with ne and ceB1 — Use ne and ce to identify a focused constituent while keeping gender, number, discourse function, and relative TAM distinct.
- Relative Completive after Object Wh-WordsB2 — Use Hausa relative-completive subject markers after fronted object questions while preserving subject agreement and distinguishing subject questions.
- Relative Continuative after Object Wh-WordsB2 — Use Hausa nake, kake, kike, yake, take, muke, kuke, suke, and ake when a fronted wh-expression asks for an object in an ongoing, habitual, or state-like clause.
- Subject FocusB1 — Identify or correct the subject of a Hausa clause through focus particles, relative TAM, and discourse-appropriate subject marking.