Proceedings of Recent Advances in Natural Language Processing, pages 839–846, Varna, Bulgaria, Sep 4–6 2017. https://doi.org/10.26615/978-954-452-049-6_107 Using NLP for Enhancing Second Language Acquisition Leonardo Zilio Rodrigo Wilkens CENTAL Universit´ e catholique de Louvain {leonardo.zilio,rodrigo.wilkens,cedrick.fairon}@uclouvain.be C´ edrick Fairon Abstract This study presents SMILLE, a system that draws on the Noticing Hypothesis and on input enhancements, addressing the lack of salience of grammatical infor- mation in online documents chosen by a given user. By means of input enhance- ments, the system can draw the user’s at- tention to grammar, which could possi- bly lead to a higher intake per input ratio for metalinguistic information. The sys- tem receives as input an online document and submits it to a combined processing of parser and hand-written rules for detecting its grammatical structures. The input text can be freely chosen by the user, providing a more engaging experience and reflecting the user’s interests. The system can en- hance a total of 107 fine-grained types of grammatical structures that are based on the CEFR. An evaluation of some of those structures resulted in an overall precision of 87%. 1 Introduction Research on the field of second language acquisi- tion (SLA) has already shown that the mere pre- sentation of input to a language learner is not enough for ensuring that some linguistic informa- tion will be retained (Meurers et al., 2010). This means that the language learner may process the input for its meaning alone, without noticing its linguistic structures, because there is no salient grammatical information. Input is, therefore, un- derstood as “potentially processible language data which are made available, by chance or by design, to the language learner” (Smith, 1993). On the other hand, the intake is the part of the input which is actually internalized by the user and that can potentially be connected to the long-term memory (Reinders, 2012). As such, an input in its raw form has lower chances of being converted into intake by the learner, and may thus not provide any new linguis- tic information. In the early 90’s, Schmidt (1990) developed the hypothesis that, in order to convert input into intake, a language learner needs to no- tice the relevant information in the input. More recently, Schmidt (2012) stated, in a less contro- versial way, that “people learn about the things that they attend to and do not learn much about the things they do not attend to”. There is much dis- cussion regarding the assumptions of the Noticing Hypothesis, and it has some fierce contesters, such as Truscott (1998). Nevertheless, it seems to be of general agreement that noticing is at least a facili- tator of the language learning process, even though there is differences in the way that authors view the process of noticing, either as a purely con- scious process or as a possibly unconscious pro- cess (Cross, 2002). To solve the lack of salience in raw input, Smith and Truscott (2014) suggested the use of “input enhancements”, so as to give prominence to the relevant linguistic information. This focus-on- form strategy (Doughty, 1991) provides a way to assist language learners, and recent studies on SLA have shown that input enhancements repre- sent a positive step in transforming input into in- take (Plonsky and Ziegler, 2016; Simard, 2009). CALL systems that are able to deal with authen- tic texts and uses NLP for rendering a better pre- sentation of linguistic information are called Au- thentic Texts Intelligent Computer-Assisted Lan- guage Learning (ATICALL) Meurers (2012). In this paper, we present the Smart and Intelligent Language Learning Environment (SMILLE), an ATICALL system that enhances authentic Web pages by using available NLP tools. It extracts the 839