Teacher Perception of Automatically Extracted Grammar Concepts for L2 Language Learning

  • 2022-06-10 15:52:22
  • Aditi Chaudhary, Arun Sampath, Ashwin Sheshadri, Antonios Anastasopoulos, Graham Neubig
  • 4


One of the challenges of language teaching is how to organize the rulesregarding syntax, semantics, or phonology of the language in a meaningfulmanner. This not only requires pedagogical skills, but also requires a deepunderstanding of that language. While comprehensive materials to develop suchcurricula are available in English and some broadly spoken languages, for manyother languages, teachers need to manually create them in response to theirstudents' needs. This process is challenging because i) it requires that suchexperts be accessible and have the necessary resources, and ii) even if thereare such experts, describing all the intricacies of a language istime-consuming and prone to omission. In this article, we present an automaticframework that aims to facilitate this process by automatically discovering andvisualizing descriptions of different aspects of grammar. Specifically, weextract descriptions from a natural text corpus that answer questions aboutmorphosyntax (learning of word order, agreement, case marking, or wordformation) and semantics (learning of vocabulary) and show illustrativeexamples. We apply this method for teaching the Indian languages, Kannada andMarathi, which, unlike English, do not have well-developed pedagogicalresources and, therefore, are likely to benefit from this exercise. To assessthe perceived utility of the extracted material, we enlist the help of languageeducators from schools in North America who teach these languages to perform amanual evaluation. Overall, teachers find the materials to be interesting as areference material for their own lesson preparation or even for learnerevaluation.


Quick Read (beta)

loading the full paper ...