Clefts 381
In (191), A asserts that out of the candidate set, Siti is the only one who bought
a car. The presupposition in A’s statement is that someone bought a car. There-
fore, the only pragmatically consistent contradiction of this assertion is one that
includes the same presupposition, that is that someone bought a car. For that
reason, only B1, which continues this presupposition but corrects the focused
set, is acceptable in the discourse. B2 maintains the focus set, Siti, but changes
the presupposition (that is that someone bought a boat not that someone bought
a car) and is thus pragmatically odd. B3 is pragmatically odd for the same rea-
son, but also changes the focus set from Siti to Ali. The simple declarative
clause in B4, that Siti bought a boat, is pragmatically well-formed as it does not
contradict the presupposition that someone bought a car. The use of banne
simply asserts that Siti, in fact, bought a boat and pragmatically entails that she
did not buy a car, not that no car was purchased. If Siti had, in fact, bought the
boat as well, banne would be pragmatically odd here.
The data in (192) illustrate the same effect.
(192) A: Buku se e-kerem Siti dhari Amerika.
book REL OV-send Siti from America
‘A book is what Siti sent from America.’
B1:Banne, koran se e-kerem Siti dhari Amerika.
no newspaper REL OV-send Siti from America
‘No, a newspaper is what Siti sent from America.’
B2:#Banne, buku se e-kerem Adi dhari Amerika.
no book REL OV-send A from America
‘No, a book is what Adi sent from America.’
B3:#Banne, buku se e-kerem Siti dhari Kanada.
no book REL OV-send Siti from Canada
‘No, a book is what Siti sent from Canada.’
Again, the presupposition must remain intact for pragmatic coherence. Only the
focus set can be contradicted with a cleft construction. The presupposition in
A’s statement is that Siti sent something from America, and only the structure in
B1 maintains that presupposition.
9.2 Predicative nature of the clefted element
Analyses of clefts in closely related languages take the focused element to be
the predicate (see Cole, Hermon, and Aman (2008) for Singapore Malay, Sned-
don (1996) for Indonesian, Paul (2001) for Malagasy). Evidence for the cor-