A Hausa dependency is long-distance when a phrase at the left edge receives its grammatical role inside a lower clause. In Wa ka ce yaran suka gani? “Whom did you say the children saw?”, initial wa is not an argument of ce “say”; it is understood as the object of gani in the embedded clause. The words between the wh-expression and that role do not cancel the connection, but clause type and argument structure constrain it.
All unlabelled examples are (neutral Nigerian Standard Hausa). Matrix clause, embedded clause, bridge verb, gap, and resumptive pronoun are (academic) terms. Ordinary boko keeps hooked letters and apostrophes but omits tone and vowel length. Some advanced judgments vary with speaker, dialect, and discourse setup; the productive models below use Nigerian Standard bridge complements and reserve stronger island claims for Island Constraints in Questions and Relatives.
From two complete clauses to one dependency
Begin with a neutral embedded statement:
Ka ce yaran sun ga Musa a kasuwa.
You said that the children saw Musa at the market. (neutral matrix plus embedded S–TAM–V–O)
The matrix clause is ka ce “you said”; the embedded clause is yaran sun ga Musa. To ask for the embedded object, Hausa can place wa at the beginning:
Wa ka ce yaran suka gani a kasuwa?
Whom did you say the children saw at the market?
The object after gani is absent. Plural suka, rather than plain sun, makes the embedded focus/relative environment visible. Matrix ka ce retains its own second-person completive marking. English inserts did in its matrix question; Hausa does not.
Subject and object dependencies
A fronted phrase may correspond to an embedded subject:
Wa ka ce ya buɗe akwatin?
Who did you say opened the box?
The embedded ya supplies masculine singular subject agreement and relative completive TAM. It is not an object pronoun after ce. Ordinary boko spells this ya like plain completive ya, so a plural example exposes the structure more clearly:
Su wa ka ce suka ɗauki akwatunan?
Who all did you say took the boxes?
For an embedded object, the lower subject remains overt:
Wane littafi ka ce ɗaliban suka karanta?
Which book did you say the students read?
Ɗaliban controls lower suka; wane littafi is interpreted after karanta. This is not free scrambling across two clauses. The fronted interrogative, the lower relative TAM, and the gap form a single licensed question dependency.
With future za, there is no separate relative future paradigm to display:
Wa ka ce za su tura Kano gobe?
Whom did you say they would send to Kano tomorrow?
The lack of a visible plain/relative contrast after za does not make the dependency local or unstructured.
Relative clauses can also be long-distance
A relative head may receive its role inside a complement:
Ga mutumin da na ce yaran suka gani a asibiti.
Here is the man whom I said the children saw at the hospital.
The relative head mutumin is interpreted as the object of lower gani, not as an object of matrix ce. The outer relative marker da opens the relative domain, and lower plural suka marks the clause containing the gap.
Oblique relations normally keep their required preposition and a resumptive pronoun:
Ga mutumin da na ce Amina ta yi magana da shi.
Here is the man whom I said Amina spoke with.
The phrase da shi preserves the complement of da “with.” English can strand with at the end while leaving only a gap; Hausa’s reliable strategy keeps da with an overt pronoun. The pronoun agrees with the masculine relative head.
Possessors likewise require a visible link rather than a bare gap. Because such long dependencies are cognitively heavy, everyday speakers often repeat a noun or divide the message into two sentences. That is a clarity choice, not grammatical failure.
Bridge predicates and scope
Productive long dependencies are clearest with predicates such as ce “say” and tsammani “think/suppose.” Bridgehood is lexically sensitive in Hausa: do not assume that every speech, cognition, factive, or manner predicate permits extraction merely because its English translation does.
Me kuke tsammani ma’aikatan suke gyarawa?
What do you all think the employees are repairing?
The matrix continuative kuke tsammani agrees with “you all.” The embedded relative continuative suke gyarawa agrees with ma’aikatan and licenses the object gap. Each clause retains a subject–TAM layer.
Compare an embedded question whose wh-expression has only local scope:
Ban tuna ko wa yaran suka gani ba.
I do not remember whom the children saw.
Here ko wa belongs inside the complement of tuna. The speaker makes a statement of ignorance; the whole utterance is not a direct question. Moving a phrase across an already interrogative ko boundary is not the ordinary long-distance pattern.
Negation belongs to the clause it semantically negates
Matrix negation can deny knowledge or a report while leaving the lower proposition as the content under discussion:
Ban san ko wa yaran suka gani ba.
I do not know whom the children saw.
Lower negation instead denies the embedded event:
Wa ka ce yaran ba su gani ba?
Whom did you say the children did not see?
The embedded completed negative ba su gani ba remains intact. The fronted wh-expression asks for its missing object. Do not wrap the matrix and lower predicates in one giant ba ... ba frame; each finite domain has its own subject marking and polarity.
A polar question can question the matrix speech event while leaving a complete complement:
Ka ce yaran sun ga Musa?
Did you say that the children saw Musa?
This asks whether the addressee said the proposition. Wa ka ce yaran suka gani? instead requests the identity of an embedded participant. English auxiliary order distinguishes the two; Hausa relies on the wh-expression, clause structure, and relevant TAM forms.
Topic chains are not extraction chains
Long-distance topicalization can avoid a compact gap by keeping reference overt inside the lower clause:
Musa kuwa, na ji cewa yaran sun gan shi a kasuwa.
As for Musa, I heard that the children saw him at the market.
Musa kuwa is a detached topic, not a matrix wh/focus phrase. The complement retains plain sun, and shi resumes Musa as the object of gan. Compare the focused question Wa ka ce yaran suka gani?, whose direct-object dependency has a gap and relative suka.
This topic strategy is often clearer in speech when several clauses intervene. It also shows why “a pronoun appears somewhere below” is not enough to diagnose focus, wh-movement, or relative extraction; discourse function and TAM must agree with the analysis.
A short discourse
—Ka ce yaran sun ga wani a kusa da shagon. Wa ka ce suka gani? —Musa ne na ce suka gani. —Amma Musa kuwa, na ji cewa ba su gan shi ba. —To, wa kake tsammani suka gani? —Mai gadi ne, amma zan sake tambayarsu.
—You said the children saw someone near the shop. Whom did you say they saw? —It was Musa whom I said they saw. —But as for Musa, I heard that they did not see him. —Then whom do you think they saw? —The guard, but I will ask them again.
The first and fourth questions connect a matrix interrogative with an embedded object gap. The middle Musa kuwa instead establishes a topic and resumes it with shi inside an embedded negative clause.
Common Mistakes
1. Adding English do-support
❌ Wa did ka ce ma’aikatan suka kira?
Incorrect English calque; Hausa does not insert did.
✅ Wa ka ce ma’aikatan suka kira?
Whom did you say the employees called?
2. Keeping plain plural TAM in the lower gap clause
❌ Wace mota ka ce yaran sun gyara?
Incorrect canonical long-distance object question; sun is plain completive in the clause containing the gap.
✅ Wace mota ka ce yaran suka gyara?
Which car did you say the children repaired?
3. Omitting the lower subject–TAM marker
❌ Wa ka ce ɗaliban gani a makaranta?
Incorrect—the embedded lexical subject does not replace its subject–TAM marker.
✅ Wa ka ce ɗaliban suka gani a makaranta?
Whom did you say the students saw at school?
4. Stranding a required preposition
❌ Ga matar da na ce Hauwa ta yi tafiya da.
Incorrect bare preposition stranding in the intended relative dependency.
✅ Ga matar da na ce Hauwa ta yi tafiya da ita.
Here is the woman whom I said Hauwa travelled with.
5. Treating an embedded question as a direct matrix question
❌ Ban san ko me ma’aikatan suka gyara ba?
Wrong punctuation for the intended declarative ‘I do not know...’; the embedded question does not automatically give the whole sentence question force.
✅ Ban san ko me ma’aikatan suka gyara ba.
I do not know what the employees repaired.
Key Takeaways
- A long-distance phrase receives its semantic role from a verb or preposition inside an embedded clause.
- Each clause keeps its own subject–TAM marking; English do-support and inversion have no place in the Hausa pattern.
- Gap-bearing embedded questions and focus structures select relative TAM where Hausa has a contrasting paradigm.
- Oblique and possessive relations generally preserve a preposition, linker, or resumptive pronoun.
- Bridge complements permit productive long dependencies, but embedded questions, relatives, adjuncts, and coordinations impose additional locality constraints.
Related Topics
- Topicalization without FocusC1 — Place an established referent at the left edge as a Hausa topic while retaining ordinary TAM and resuming its role inside the comment.
- Island Constraints in Questions and RelativesC1 — Recognize the Hausa clause types that block gap-based extraction and the narrower resumptive strategy available to some relative dependencies.
- Wh-Scope in Embedded QuestionsC1 — Keep Hausa wh-expressions inside the question they belong to, using ko and relative TAM to distinguish embedded information gaps from matrix questions and focused assertions.
- Speech Verb ComplementsB1 — Combine Hausa speech verbs with addressees, quotations, factual reports, and reported directives without dropping clause-level marking.
- Object Relative ClausesA2 — Build Hausa relatives whose linked noun head is understood as the missing object of the modifying clause.
- Resumptive Pronouns in RelativesB1 — Know when a Hausa relative leaves a gap and when a pronoun or possessive suffix must preserve the head noun's grammatical relation.