On the interplay between information structure and syntax in the history of Englishan experiment with automatic co-reference resolution

Resumo

As part of a larger project on the interplay between information structure and syntax in the history of the English language, this pilot study aims at exploring the connection between word order and information status by analysing the syntactic and informative features of sentence-initial subjects and postverbal objects in the history of English. The research periods are Early Modern (1500-1710) and (Late) Modern (1700-1914) English, that is, after the syntacticisation of word order in English. The approach is corpus-based because, on the one hand, the data are retrieved from electronic collections of Early Modern and Late Modern English texts (the Penn-Helsinki Parsed Corpus of Early Modern English or PPCEME - Kroch et al. 2004 -, and the Penn Parsed Corpus of Modern British English or PPCMBE - Kroch et al. 2010 -, respectively). On the other hand, this research utilises the CESAX framework, developed at Radboud University (Komen, 2009, 2011a), which adds co-reference markup to the corpora. In other words, the software detects chains or links between subjects/objects and their potential co-textual antecedents. On theoretical grounds, the system relies on Centering and Optimality approaches, which are described in this study. By using CESAX, I will show that the referential links provided by the software do not pave the way for the linkingfunction which is attributed to sentence-initial subjects when the language is subject-prominent.