Towards Fine-Grained Readability Measures for Self-Directed Language Learning
Lisa Beinborn, Torsten Zesch, Iryna Gurevych · TUbilio (Technical University of Darmstadt) · 2012
In this paper, we analyze existing readability measures regarding their applicability to self-directed language learning. We identify a set of dimensions for text complexity and focus on the lexical, syntactic, semantic, and discourse dimensions. We argue that for the purposes of self-directed language learning, the assessment according to the individual dimensions should be preferred over the overall readability prediction. Furthermore, due to the heterogeneity of the learners in such a setting, modeling the background knowledge of the learner becomes a critical step.