Processing Hungarian: The role of topic and focus in language comprehension
This work examines the role of topic and focus in sentence processing, with particular attention to Hungarian. It is proposed that sentence topic and contrastive focus have a privileged relation to any prior discourse-related phrases, allowing them to influence the processing of the current sentence. The dissertation describes a series of experiments designed to test a distinction in processing of the discourse-related phrases topic and focus on the one hand, and non-discourse-related (background) constituents on the other. I argue that topic and focus receive discourse-interpretation during their initial processing; by contrast, elements in the background (other than inherently discourse-related ones) are incorporated into the discourse only after the entire proposition has been processed. This has important consequences for the types of information used by the parser when processing topic and focus vs. background. In particular, I propose that topic plays a distinguished role in constructing the contrast set for a contrastively focused item.
The privileged role of topic and focus is argued to follow from a syntactic property of Hungarian, namely, the structural marking of topic and focus. This allows the parser to identify the discourse-related status of topic and focus as soon as they are encountered. Immediate discourse interpretation of topic and focus is further related to the possibility of multiple filler-gap dependencies in Hungarian, which raises special issues in processing complexity and memory limitations. I examine some grammatical and parsing principles necessary to interpret simultaneous dependencies between two topics or wh-phrases and their traces. I propose that the parser constructs temporary memory buffers on the spot for each type of A$\sp\prime$-dependency (topic, wh-phrase) separately, to help relieve the burden on immediate memory posed by a series of unassigned fillers.
0633: Cognitive therapy