EasyEnglishAnalyzer: Taking Controlled Language from Sentence to Discourse Level

Controlled Language checking has traditionally been applied largely on a sentence level by placing restrictions on permissible vocabulary and permissible syntactic constructions, including proper punctuation. Only little attention has been paid to document or discourse level checking. In this paper, we report on work on EasyEnglishAnalyzer to handle certain discourse and document level checks. This work helps take Controlled Language checkers from the sentence level to the discourse and document level. Deep semantic analysis provided by a discourse understanding system assists with the semantically more challenging tasks such as proper paragraph structure. For checks related to overall style and organization, document structure is recognized by enhanced interpretation and use of document structure tags so that appropriate checks can be applied in a context-sensitive manner. Parsing of combined segments is applied in checking correctness of list environments.