Processing temporal discourse relations: why does overt linguistic marking become redundant
Ref study: 13378