Complexity of Russian Academic Texts as the Function of Syntactic Parameters

Abstract

Providing students with academic materials of steadily increasing complexity is equally critical for textbook publishers, educators, and test developers, therefore automation of text complexity assessment is relevant in a number of areas. The authors propose an innovative approach addressing limitations of the existing Russian text complexity tools based on descriptive and lexical features only. We examined the impact of 75 syntactic features on texts measured with ETAP-3, contrasted them in texts of different grade levels in Russian Readability Corpus of 1.2 million tokens and suggest an original model of text complexity. We confirm statistically significant correlation of text complexity and 28 syntactic features in Russian texts. More research is needed to examine the model’s accuracy in texts of different types and genres as well as its applicability in automated text analyzers.