On the performance of modern digital grammar checkers for native and second language learners

Marco Bril

doi:10.1075/dujal.18009.bri

Abstract

AbstractIn this study we conducted an experiment aiming to compare the performance of language learners and digital grammar checkers at supplying gender marking in French. A set of test items exhibiting typical gender marking configurations was submitted to three grammar checkers for French (Antidote, Scribens and BonPatron). The outcomes were compared to those of native speakers and second language learners of French at a B1 level.The results revealed that only Antidote and Scribens outperformed both native speakers and second language learners of French in adjective-noun and fronted noun-past participle agreement constructions. An opposite pattern, however, appeared for clitic-past participle constructions for which native speakers outperformed Antidote and Scribens.We thus conclude that from the three grammar checkers under investigation, Antidote and Scribens might be effective to improve the native speakers’ and second language learners’ awareness of gender marking errors, but only in adjective-noun and fronted noun-past participle agreement constructions.

Highlights

In the last decades, many studies aimed at evaluating the effectiveness and the usefulness of digital grammar checkers in computer-assisted language learning (e.g. Biesemans, 2005; Burston, 1998; Heift & Schulze, 2007; Vernon, 2000). Heift and Schulze (2007) for instance, showed that grammar checkers are very effective in form-focused writing exercises
We report the results of an experiment investigating the supply of gender marking in different types of gender agreement configurations by three digital grammar checkers (Antidote, Scribens and BonPatron) and by two populations of French language users
In this study we conducted an experiment aiming to compare the performance of human language users and digital grammar checkers at supplying gender marking in French

Summary

Introduction

Many studies aimed at evaluating the effectiveness and the usefulness of digital grammar checkers in computer-assisted language learning (e.g. Biesemans, 2005; Burston, 1998; Heift & Schulze, 2007; Vernon, 2000). Heift and Schulze (2007) for instance, showed that grammar checkers are very effective in form-focused writing exercises. Instruction is a pedagogical practice undertaken by second language teachers by which the students’ attention is drawn to language form In this particular case, form-focused exercises would explicitly target overt grammatical inflection by confronting students with the relevant morphosyntactic rules. Within the context of gender inflection for example, these exercises may consist of fill-in-the-blank sentences in which learners need to fill in the correct gender inflection or sentences containing incorrect gender marking which need to be corrected by the language learners. In this type of writing exercises, grammar checkers can be of help for learners.

Objectives

Methods

Results

Discussion

Conclusion