Analyzing Culture-Specific Argument Structures in Learner Essays.

Wei-Fan Chen,Mei-Hua Chen,Garima Mudgal,Henning Wachsmuth

2022-01-01

Abstract:Language education has been shown to benefit from computational argumentation, for example, from methods that assess quality dimensions of language learners’ argumentative essays, such as their organization and argument strength. So far, however, little attention has been paid to cultural differences in learners’ argument structures originating from different origins and language capabilities. This paper extends prior studies of learner argumentation by analyzing differences in the argument structure of essays from culturally diverse learners. Based on the ICLE corpus containing essays written by English learners of 16 different mother tongues, we train natural language processing models to mine argumentative discourse units (ADUs) as well as to assess the essays’ quality in terms of organization and argument strength. The extracted ADUs and the predicted quality scores enable us to look into the similarities and differences of essay argumentation across different English learners. In particular, we analyze the ADUs from learners with different mother tongues, different levels of arguing proficiency, and different context cultures.