TextMix: using NLP and APIs to generate chunked sentence scramble tasks

A Computer-Assisted Language Learning (CALL) application, TextMix, was developed as a proof-of-concept for applying Natural Language Processing (NLP) sentence chunking techniques to creating ‘sentence scramble’ learning tasks. TextMix addresses limitations of existing applications for creating sentence scrambles by using NLP to parse and scramble syntactic components of sentences, while connecting with Application Programming Interfaces (APIs) to provide repeated exposure to authentic sentences in the context of texts such as Wikipedia articles. In addition to identifying a novel application of NLP and APIs in CALL, this project highlights the need for teacher-friendly interfaces that prioritize pedagogically useful ways of chunking text.