Preserving Discourse Structure when Simplifying Text

Text simplification involves restructuring sentences by replacing particular syntactic constructs (like embedded clauses and appositives). The aim is to make the text easier to read for some target group (like aphasics and people with low reading ages) or easier to process by some program (like a parser or machine translation system). However, sentencelevel syntactic restructuring can wreak havoc with the discourse structure of a text, actually making it harder to comprehend, and possibly even altering its meaning. In this paper, we present and evaluate techniques for detecting and correcting disruptions in discourse structure caused by syntactic restructuring. In particular, we look at the issues of preserving the rhetorical relationships between the original clauses and phrases and preserving the anaphoric link structure of the text.