Complement Clauses

From Sereer wiki
Jump to navigationJump to search

Finite Complement Clauses with (y)ee/ndax

Complement clauses in Sereer generally involve the linking of the main clause with the subordinate clause using the complementizer "yee". The unmarked order appears to be the main clause followed by the subordinate clause, but inversion of this order is possible, as well.

Direct Reported Speech Pause between 1st and 2nd clause. (1) Alaya, abuga ombec. He said, "he likes to dance". (2) Alayaxam, bugaam ombec He told me, "I like to dance".

Indirect Reported Speech No pause necessary, complementizer is used. (3) Alaya yee axe fecaa. He said that he's dancing.

(4) Alaya yee mexe fecaa. He said that I'm dancing.

(5) Layam mee awe mbecaa. I said that they are dancing

Complements of belief and knowledge: same as with indirect reported speech (6) agima yee ɗaaneegaam He believes that I was sleeping

(7) afoga yee mexe ɗaanaa. He assumes that I'm sleeping.

(8) agimaɗaara yee mexe ɗaanaa. He doubts that I'm sleeping.

(9) a'anda yee mexe ɗaanaa. He knows that I'm sleeping.

(10) aweƈa yee mexe ɗaanaa. He forgot that I'm sleeping.

Non-Finite Complement Clauses with o

Non-finite clauses in Sereer are formed with the preverbal marker o and a bare verb form, either with no final vowel or a final -aa suffix if imperfective. Non-finite clauses are found as the complement to control verbs when the subject of both clauses is coreferential. So far these verbs include:

  • bug want to, like to
  • war must
  • ...

Structurally, these clauses appear to be much smaller than other types of complement clauses, this is because:


  • They allow no subject agreement morphology
  • They allow no tense morphology
  • They may not preceded by the complementizers (y)ee or ndax
  • There is no focus or wh-word position available before the embedded verb. Any extraction element from the o-clause must precede the matrix verb
  • They do not allow the finiteness marker -a or the extraction marker -u


The bond between the matrix and embedded verb heading a non-finite o-clause seems to be much tighter than in other constructions. Two phenomena point to this conclusion. First, an object clitic associated with thematically with the lower verb may be attached to the matrix verb. This cannot be analyzed as fronting in the lower clause, as full noun phrases may not intervene between the higher and lower verb.


Second, when extracted, certain adjunct types require the presence of an applicative suffix on the verb of their clause when they are extracted. When adjuncts are extracted out of a non-finite complement clause, the applicative suffix may appear on (1) only the embedded verb; (2) both the embedded verb and the matrix verb; or (3) just the matrix verb.

<gl fontsize=11> tam Jegan abugu (o) gimit \gll tam Jegan a= bug-u (o) gim-it where Jegan 3sg.sbj want-foc inf sing-appl \trans Where does Jegan want to sing?. </gl>

<gl fontsize=11> tam Jegan abugtu (o) gimit \gll tam Jegan a= bug-t-u (o) gim-it where Jegan 3sg.sbj want-appl-foc inf sing-appl \trans Where does Jegan want to sing?. </gl>

<gl fontsize=11> tam Jegan abugtu (o) gim \gll tam Jegan a= bug-t-u (o) gim where Jegan 3sg.sbj want-appl-foc inf sing-appl \trans Where does Jegan want to sing?. </gl>

So, the matrix verb and the embedded verb seem to behave as a single predicate for purposes of object clitic placement and applicative marking with triggered by extraction. Neither is possible with other complement clause types.

Factive Clauses

Factive clauses are used as subject and complements of predicates that presuppose the truth of their subject or complements. In Sereer, these clauses are morphosyntactically distinct from finite embedded clauses. They are formed with a headless relative clause.