Evaluating Lexical Simplification and Vocabulary Knowledge for Learners of French: Possibilities of Using the FLELex Resource

Evaluating Lexical Simplification and Vocabulary Knowledge for Learners of French: Possibilities of Using the FLELex Resource Ana¨ıs Tack1, Thomas Franc¸ois1,3, Anne-Laure Ligozat2, C´edrick Fairon1 1Centre for Natural Language Processing (CENTAL), Institute for Language & Communication (IL&C), Place Blaise Pascal 1, University of Louvain, B-1348 Louvain-la-Neuve, Belgium 2LIMSI, CNRS and ENSIIE, rue John Von Neumann, 91400 Orsay, France 3FNRS Postdoctoral Researcher anais.tack@uclouvain.be, thomas.francois@uclouvain.be, annlor@limsi.fr, cedrick.fairon@uclouvain.be Abstract This study examines two possibilities of using the FLELex graded lexicon for the automated assessment of text complexity in French as a foreign language learning. From the lexical frequency distributions described in FLELex, we derive a single level of difficulty for each word in a parallel corpus of original and simplified texts. We then use this data to automatically address the lexical complexity of texts in two ways. On the one hand, we evaluate the degree of lexical simplification in manually simplified texts with respect to their original version. Our results show a significant simplification effect, both in the case of French narratives simplified for non-native readers and in the case of simplified Wikipedia texts. On the other hand, we define a predictive model which identifies the number of words in a text that are expected to be known at a particular learning level. We assess the accuracy with which these predictions are able to capture actual word knowledge as reported by Dutch-speaking learners of French. Our study shows that although the predictions seem relatively accurate in general (87.4% to 92.3%), they do not yet seem to cover the learners’ lack of knowledge very well. Keywords: FLELex, lexical simplification, vocabulary knowledge prediction 1. Introduction In the area of second language acquisition, it is argued that mastering vocabulary is essential for skilfully practising a second or foreign language (L2). For L2 reading in par- ticular, a number of studies have stressed the importance of knowing a sufficient amount of words in a text – at least 95% of a text’s vocabulary to be precise – in order to achieve adequate reading comprehension (Laufer and Ravenhorst-Kalovski, 2010). It is therefore important to present a learner with suitable reading material, which in- cludes vocabulary that is tailored to the learner’s level of proficiency. In this view, recent advances in natural lan- guage processing (NLP) have led to the development of tools that automatically adapt the textual content to the learner (Burstein, 2009). One way such content adaptation could be achieved is via automatic text simplification (ATS), as it alters the text to adjust its difficulty to the learner’s level. ATS is commonly carried out through syntactic transformations (e.g. sentence splitting, sentence deletion, etc.) and lexical substitutions. Both can be induced from a parallel corpus of original and simplified texts (Woodsend and Lapata, 2011). How- ever, the availability of such corpora is an issue, espe- cially for French and even more for French L2. Indeed, the commonly used corpus constitutes an aligned version of Wikipedia and Simple Wikipedia (Zhu et al., 2010) and is thus mainly intended for native …

Content truncated to 500 words