AZCALL 2020 

Virtual Conference

From Corpus to Classroom: Evaluating Web-based Tools to Teach Collocation  

Larissa Goulart & Maria Kostromitina; Northern Arizona University

Contact: lg845@nau.edu 

Biographies

Larissa Goulart is a PhD Candidate in Applied Linguistics at Northern Arizona University. Her research focuses on register variation in second language writing for university purposes.

Maria Kostromitina is a PhD student in Applied Linguistics at Northern Arizona University. Her research interests are L2 Psycholinguistics, especially speech parsing and L1 transfer and L2 pragmatics. 

*Ask questions and comment below


Abstract

Nesselhauf (2005) defined collocations as syntagmatic relations between words, that is, collocations are successions of linguistic entities that are processed as an integral unit  (Durrant & Mathews-Aydinli, 2011:60). Several studies (Chang et al. 2008; Schmitt, 2012) have already discussed the importance of the fluent use of collocations by English Language Learners (ELLs). These studies suggest that learning collocations is important as they make a learner’s text or speech more fluent and indicate belonging to a specific speech community. Cobb (2010) proposes that, since collocations are identified with the use of computer tools, ELLs should also use automated tools to learn them. These Web-based Learning Tools (WBLT) should give learners resources extracted from corpus linguistics research that helps them achieve an accurate and fluent production of collocations. Taking this into consideration, this study seeks to evaluate 5 WBLTs that aim at helping L2 learners of English produce accurate collocations, from a researcher and teaching perspective: Flax, SkELL, Linggle, Just the Word, Netspeak. This evaluation was divided in 3 parts: 1) the research conducted in the development of the WBTL,  2) the WBLTs design and accessibility  and 3) the WBLT pedagogical applications. The results of our study show that most of these tools rely on frequency-based collocations. In addition, each tool contributes to different types of class activities. For instance, Flax emphasizes the role of genre variation, while Linggle highlights collocations with different parts-of-speech. Considering pedagogical implications, we have noticed that these tools vary to a great extent in the collocation search: while Linggle requires some in-class training before learners can use these independently, SkeLL is self-intuitive. We conclude this presentation with task proposals that could be implemented in the ELL classroom using these WBLT tools to teach collocations.

Goulart & Kostromitina - Comments and Questions (Responses)