On the Influence of Text Complexity on Discourse-Level Choices

Text complexity can be reduced by making different choices at the lexical and grammatical levels. However, discourse-level choices may also affect a text’s complexity. In a coherent text, explicit discourse relations (e.g. CAUSE, CONDITION) are expressed using discourse markers (e.g. since, because, etc.) that may be preferred for texts at different readability levels. In this paper, we investigate the differences in discourse properties of texts across readability levels. In particular, we investigate the effect of readability level on (1) the usage of discourse relations, (2) the usage of discourse markers and (3) the distribution of discourse markers signaling explicit discourse relations. Our analysis of the Simple English Wikipedia corpus shows that complex and simple texts seem to have the same distribution of discourse relations; however, these relations are expressed using different discourse markers depending on the readability level of the text.