Abstract

This paper argues that focus-ground does not necessarily determine the quantificational structure of focus-sensitive operators. It shows that these operators may express their semantics on partitions other than the focus-ground partition. This means that recursive or overlapping focus-ground partitions are not required in sentences with more than one focus-sensitive operator. The belief that more than one focus-ground partition per sentence may be available appears to rest in part on the assumption that every pitch accent is correlated with a focus in a focus-ground partition. Since sentences may have more than one pitch accent, that means they contain more than one focus. This assumption, however, is unwarranted. Accenting is used as a structural resource in natural language for a number of different reasons. It is shown below that there is no one-to-one correspondence between accent and focus and between accent and an `operator-associated' element.

Citations