Using linguistically motivated features for paragraph boundary identification