Factive Relative Clauses in Wolof

In this talk, I investigate factive constructions in Wolof, an Atlantic language of Senegal.  I show that the constructions, the “li-factive” (1), “verbal factive” (2), are formally relative clauses and I argue that they all involve null nouns in SpecCP.  In the li-factive (1), a relative clause complementizer, l-i, occurs on the left edge of the (bracketed) clause.  In contrast, the verbal factive (2) involves a copy of the verb preceding a different relative clause complementizer (b-i):

(1)  Réccu-naa [l-i           ma  jaay  tééré  b-i  ].                           li-Factive

regret-1sg  cl-C_rel  1sg  sell   book  cl-the

‘I regret (the fact) that I sold the book.’

(2)  Réccu-naa [jaay [b-i        ma   jaay  tééré   b-i]].                  Verbal Factive

regret-1sg  sell   cl-c_rel  1sg   sell    book  cl-the

‘I regret (the fact) that I sold the book.’

‘I regret the way that I sold the book.’

I argue that the factive constructions in (1-2), ‘factive relatives’, involve two distinct null fact nouns that trigger (noun class) agreement on the relative clause complementizers (l‑i and b-i). Using evidence from copied material, I further argue that the null fact noun in the verbal factive originates in the TP region, lower than T^0, and undergoes A¢-movement to SpecCP. The null noun pied pipes lower material, which may or may not be pronounced.  The Wolof data provides strong cross‑linguistic support for analyses that treat factive clauses as involving operator movement (Aboh 2005, Haegeman 2006) into the left periphery of the clause Finally, I also compare and contrast the Wolof factive relative constructions to similar clause types in other languages, such as Pulaar, a closely related Atlantic language.

Harold Torrence is an associate professor of linguistics at the University of California Los Angeles. His research focuses on the syntax and morphology of African and Native American languages.