The interpretation and realization of focus: An experimental investigation of focus in English and Hungarian

Linguists have associated the word focus with many types of phenomena. In order to investigate one type of focus one must control for the effects of the others. That is the core of this study: investigating which interpretation is associated with which realization once the effects of other types of focus have been factored out. The Animation Corpus was collected to provide data in which context was controlled so that three types of focus, sc INFORMATION STRUCTURE FOCUS, sc DISCOURSE NEW/ sc OLD and sc RELEVANT NON- sc UNIQUENESS, could be separated and reliably coded. The Animation Corpus also provided data that was comparable for English and Hungarian since utterances were collected using the same animated context and tasks. There are four major results from this work. First this study demonstrates more convincingly than previous work that in English, entire constituents are marked as sc INFORMATION STRUCTURE FOCUS by duration and amplitude. This differs from a very common view that marking one word within a constituent serves to mark the whole constituent as focus. The second major result is showing that current syntactic accounts of focus in Hungarian do not cover all the patterns found in the Animation Corpus. I suggest that metrical constraints are an additional factor to consider in accounting for the distribution of sc INFORMATION STRUCTURE FOCUS. Another contribution of this study is the development of techniques used in collecting the Animation Corpus. These techniques, using animations and drawings in a task, succeeded in eliciting corpora in both English and Hungarian that had useful control on discourse factors and vocabulary but remained relatively spontaneous. A fourth contribution of the study is in the use of matched pairs for analysis. Matched pairs of items, in this study words, share the same values for all relevant properties except the one being investigated. My use of this pairing facilitates the isolation of factors that could affect prosody and allows me to demonstrate that sc DISCOURSE NEWNESS is not realized prosodically in English