Evaluating Classroom Potential for Card-it: Digital Flashcards for Studying and Learning Italian Morphology

Mariana Shimabukuro, Jessica Zipf, Shawn Yama, Christopher CollinsACL 2023

TL;DR

This paper introduces "Card-it," a web-based application designed to help learners master Italian verb conjugation through digital flashcards. By integrating a sophisticated morphological analyzer with a user-friendly interface and classroom management tools, Card-it aims to bridge the gap between theoretical linguistic knowledge and practical application for second language learners.

What Problem Does This Paper Address?

Learning the morphology of highly inflected languages like Italian is a significant challenge for second language (L2) learners. Traditional methods often rely on rote memorization, which can be tedious and fail to foster deeper understanding. While Natural Language Processing (NLP) tools, such as finite-state morphological analyzers, can provide detailed linguistic insights, they are often not designed for direct use by language learners, lacking intuitive interfaces and pedagogical features. This paper addresses the need for a tool that makes advanced NLP capabilities accessible and engaging for L2 learners, specifically focusing on Italian verb conjugation, and explores its potential for classroom integration.

Methods

Card-it is built around a Finite-State Morphological (FSM) analyzer that can both analyze existing verb forms and generate new ones based on linguistic rules. This FSM contains a lexicon of over 5000 Italian verb lemmata and their inflectional paradigms. The system leverages this FSM in a semi-automatic flashcard generation process.

The core innovation lies in the integration of the FSM with a user-friendly web application. The front-end, developed with React.js, provides a flashcard interface where one side displays a verb form and the other its morphological analysis. Card-it also offers:

  • Customizable flashcard decks and collections: allowing users and teachers to organize content.
  • Multiple study and assessment modes: including flip-card review, a "Conjugate" mode where users type the correct form, and an "Identify Tense" quiz.
  • Meta-linguistic information: explanations of linguistic categories are available on demand.
  • Corrective feedback: informed by the FSM's analysis.
  • Classroom management features: enabling teachers to create classrooms, share collections, and track student progress.

The design process was iterative, incorporating feedback from preliminary expert evaluations with Italian teachers.

Datasets

The FSM lexicon was created by extracting verb roots from publicly available resources, including the Morph-it! lexicon and the online dictionary from "Corriere della Sera." The paper does not describe the creation of new datasets for training or evaluation in the traditional machine learning sense, but rather the construction of a linguistic resource (the FSM lexicon) for the application.

Key Insights and Ideas

  • Bridging NLP and L2 Learning: The paper demonstrates a successful approach to making powerful NLP tools (FSMs) directly usable and beneficial for language learners, moving beyond their typical use in backend processing pipelines.
  • Meta-linguistic Awareness is Key: Explicitly presenting morphological information and definitions of linguistic terms is crucial for fostering deeper understanding and improving L2 acquisition.
  • Iterative Design with Expert Feedback: Incorporating feedback from domain experts (Italian teachers) early and often in the development cycle is vital for creating a tool that meets pedagogical needs and has practical classroom potential.
  • Classroom Integration Potential: The implementation of classroom management features significantly enhances Card-it's utility, transforming it from a standalone learning tool into a valuable supplementary resource for teachers.

Key Results

An expert evaluation with 9 Italian teachers (5 volunteered, 3 completed) yielded positive results:

  • Usability: Rated positively, with experts finding it "easy" or "very easy" to use.
  • Quiz Modes: Both "Conjugate" and "Identify Tense" quizzes were rated as "appropriate" or "very appropriate," with one expert calling them their "favourite features."
  • Classroom Management: Usability was rated "good" or "very good."
  • Potential Use: Experts foresaw using Card-it for homework assignments and as a tool for student self-study.
  • Suggestions for Improvement: Teachers suggested adding translations and example sentences to the flashcards.

Limitations and Future Work

  • Limitations: The current evaluation was based on a small number of expert teachers, not actual L2 learners. The FSM updates are currently manual.
  • Future Work:
    • Conduct a longitudinal user study with L2 learners to gather more reliable data on effectiveness.
  • Gamification of quizzes and modes to increase engagement.
  • Explore features that can support live classroom exercises.
  • Potentially automate FSM updates.
  • Incorporate translations and example sentences as suggested by teachers.

Summary updated

BibTeX
@inproceedings{shimabukuro2023evaluating,
  author = {Mariana Shimabukuro and Jessica Zipf and Shawn Yama and Christopher Collins},
  title = {{Evaluating Classroom Potential for Card-it: Digital Flashcards for Studying and Learning Italian Morphology}},
  booktitle = {Annual Meeting of the Association for Computational Linguistics},
  year = {2023},
}
Figure