ABSTRACT:

"Canonical Information Structures"

Nomi Erteschik-Shir, Ben-Gurion University

Canonical Information Structure - the unmarked alignment of word order and information structure – plays a role in constraining both syntactic  dependencies and scope: Whereas dependencies are strongly constrained by Canonical Information Structure, scope is only weakly constrained and therefore non-canonical scopal interpretations are possible albeit marked. Since Canonical Information Structure varies cross-linguistically constraints on dependencies and scope also vary from language to language. The English canonical information structure, for example, requires isomorphism between syntax and IS with Subject, VP aligning with Topic, Focus. In Danish it suffices for the topic to be preverbal and the focus to be postverbal. The seeming absence of the superiority effect in Danish as opposed to English is shown to follow.

     The fact that object wide scope is marked in English, but not impossible, follows on the view that truth values are calculated with respect to topics which therefore take wide scope: In order for an object to take wide scope it must be interpreted as a topic, a non-canonical Information structure. In Danish, a postverbal object topic is non-canonical which predicts that object wide scope is marked as it is in English. It also follows that when the object is topicalized (in Danish) it must take wide scope.

     I explore the view that the reason canonical alignment plays such an important role in grammar is due to the fact that such isomorphism facilitates processing.