Creating a Classroom-MT: Connecting Simplification Methods to Language Learner Levels in Monolingual Machine Translation

Abstract

This study aims to investigate whether there is a theoretical basis for the creation of a Classroom-MT, that would be able to simplify high level texts to the level of an L2 language learner. This paper will classify language learners based on the CEFR levels presented by the Council of Europe (2001; 2018) and variables that determine the developmental sequence of an L2 presented by Goldschneider and DeKeyser (2001). It will then give an overview of state-of-the-art (monolingual) machine translation models. The models will be examined both amongst each other as well as their usefulness in the creation of the Classroom-MT. This study concludes that there is a stable theoretical basis to create a Classroom-MT, in both the language learner levels as well as the simplification techniques available. Future research is needed in both the search for quantifiable variables connected to the developmental sequences, as well as how this Classroom-MT can be implemented correctly.