Observations of teaching and learning with corpora

Abstract

The benefits of using corpora for language teaching have been well researched—namely, bringing authentic language into the classroom, correcting teacher intuition, and promoting autonomous learning, to name a few (O'Keeffe et al., 2007). However, the use of corpora in mainstream teaching practice remains low (Poole, 2020). Poole (2020) suggests that this is because few ready-made corpus materials exist for teachers to use in the classroom, noting that most corpus resources are not intuitive to users. This article presents an exploration of Charles's (2015) work in teaching doctoral students to use corpora, wherein she outlines how students can create their own “DIY corpora” to investigate discipline-specific academic writing. For corpora to move from theory to practice, researchers have called on language teachers to teach with corpora and report on their experiences in their classrooms (Callies, 2019). Using corpora, teachers can deepen their own language knowledge and awareness, expand their understanding of language use in different contexts, and gain insights into learners' language needs and common errors, thus enabling them to make more informed instructional decisions and design effective language teaching materials. Provided here for novice corpus teachers to demystify the process are observations I made with my students when learning to use a corpus in the classroom. Charles (2015) trained students to build their own discipline-specific corpora to promote language autonomy and awareness. I followed her design but adapted it by showing students how to build their small corpora in Sketch Engine instead of AntConc. The non–credit bearing module was offered at an Irish university and aimed to help English language learners improve their academic writing in their master's programmes. Students in this module understood academic terms such as hedging, linking devices, and transitions, but still struggled as second language writers. They enrolled in this module to help them write their master's dissertations. Most students' native language was Chinese. Charles (2015) recommends introducing students to different corpora with active and reflective tasks. As materials for this module, the tasks I adapted came from an online corpus course through the University of Sheffield. The module was carried out over 12 weeks, with each class being approximately 90 minutes. Table 1 provides the syllabus for the module. Throughout the module, we concluded three main observations: 1) acknowledge the challenges and persevere; 2) less is more; and 3) specificity is key. Learning to use a corpus can be challenging for both the teacher and the students mainly due to the amount of data (or words) in the corpus. However, we learned that corpus websites like Sketch Engine allowed us to choose how many sentences to download beforehand, which for teachers means that you can show fewer examples to students. For students, it means they can control how many samples they wanted to focus on or analyse. Another challenge we found was that few corpora were freely accessible by the public. This is because of the time and cost required to build them. Although our university has a subscription to Sketch Engine, students also requested to see another corpus for comparison at the end of the module. We found some free corpora such as SKeLL and The Corpus of Contemporary American English (COCA). COCA asks users to register and then limits the number of searches a day to 25 but this was enough for our purposes in class. The benefit of persevering through these challenges means greater authenticity in language learning. When we first began using a corpus, the task was to find 20 of the most frequently used nouns and verbs within their individual corpora. This task took most of the lesson time because both the students and I were new users to the corpus. We considered that searching for 10 nouns and verbs might have been sufficient. However, we wanted to learn as much as we could from using corpora, and practicing daily for an extended period of time could provide similar training without overwhelming students with new corpus tools. Once we became more familiar with using the tools in Sketch Engine, we soon learned how critical it is to be specific in search queries. Thankfully, Charles (2015) gives specific tasks such as looking for academic linkers such as nevertheless and nonetheless, so we had a guide but we also learned how to use wild cards to yield more accurate results for our searches. For example, a student wanted to identify the most frequent preposition with comparison from their corpus using the asterisk wildcard, typing in * comparison. Since * can replace any word (or letter), the corpus will give a ranked list of prepositions with comparison. In using the wild card and the word they were interested in, the student conducted a specific search. Sketch Engine offers video tutorials of how to perform searches when first starting out. I'm so very grateful for the sources you provided me with. I fancy the contents and the tools. Now there are always two web pages, COCA and Sketch Engine, open while I'm writing my essay. Thank you for giving us such amazing lectures. I learned a lot from the “Sketch Engine,” and the practical benefit for my dissertation is—if I need to check word usage, I can relate it directly to my dissertation corpus and find the specific journal articles. We can conclude from this experience that students found it beneficial to learn academic writing with a corpus. Although we tackled some challenges at the beginning of the process, they were outweighed by the rewards. We learned to pace ourselves when first learning to use a corpus for language learning, and that specificity in corpus searches was invaluable for obtaining the answers to our linguistic queries about word frequency and usage. There are schools which offer corpus training summer schools such as the University of Lancaster and the University of Birmingham which I have recommended students and trainee teachers to try. I hope my experience of teaching English for academic purposes (EAP) with corpora encourages more practitioners to try out corpora in their classrooms. Cathryn Bennett recently completed her PhD in the School of Speech, Linguistics and Communication Sciences at Trinity College Dublin. Her doctoral research focuses on teacher training with corpora for in-service teachers of English as a foreign language (EFL). She has been an EFL practitioner for 14 years, teaching in Europe and Asia.