Event date:

WA Lunch Talk: “Uncovering Patterns in Dictionary Look-Up Behavior: Machine Learning Meets CEFR Vocabulary Levels” and “Strategies for effective communication in Dutch as a lingua franca telecollaboration”

WA Lunch Talks logo

We are delighted to invite you to the upcoming WA Wednesday Lunch Talk, which will take place on December 11th at 13:15 in Aula Hrynakowskiego.

This session will feature two presentations:

  1. Prof. Robert Lew & Sascha Wolfer: “Uncovering Patterns in Dictionary Look-Up Behavior: Machine Learning Meets CEFR Vocabulary Levels”
  2. As part of the Wiktionary project (NCN Poland), we analyzed over a decade of user interactions with the English Wiktionary (likely the most extensive online dictionary available). We were particularly interested in how lexical variables such as corpus frequency, age of acquisition, degree of polysemy, word prevalence, part-of-speech, word-length, and subsequently CEFR vocabulary level correlate with user interest in specific entries. As a sort of bonus to the project, and something we hadn’t planned to do at the outset, we also tried training machine learning algorithms to classify vocabulary items into CEFR levels, based on lexical factors and Wiktionary views. Our results suggest that this approach is a cost-effective way to update and supplement existing CEFR-graded vocabulary lists.

  3. Prof. Aleksandra Wach, dr Robert de Louw, & dr Mikołaj Buczak: “Strategies for effective communication in Dutch as a lingua franca telecollaboration”
  4. Research into communication strategies (CS), defined as behaviors employed to overcome problems or facilitate communication (Dörnyei & Scott, 1997), has been recently reinvigorated due to developments in computer-mediated communication (CMC) and its growing role in L2 instruction. Studies on CS in telecollaboration, which is educational practice engaging groups of students in collaborative tasks with peers from different cultural backgrounds (O’Dowd, 2018), have investigated the impact of CS use on learning opportunities, pointing to its variability and specificity regarding the context (Cirit-Işıklıgil et al., 2023; Hung & Higgins, 2016). However, research into lingua franca telecollaboration, especially in less commonly taught languages (LCTL), has been largely under-represented in the literature (Çiftçi & Savaş, 2018; Godwin-Jones, 2019). Our study thus attempts to contribute to this research gap by investigating CS use by Polish and Hungarian students using Dutch as a lingua franca in video-based synchronous computer-mediated communication (SCMC). The data were collected through recordings of 12 videoconferences held by four Polish-Hungarian groups (N=21) and through post-exchange interviews conducted with the Polish students (n=13). The data underwent a qualitative content analysis employing the Atlas.ti software with the aim of tracing the use of CS during the interactions and the participants’ reflections about the CS they used. The findings revealed a variety of avoidance, compensatory, interactional, and paralinguistic strategies, with fillers, non-verbal strategies, and repair being the most frequent CS. Interestingly, a high proportion of CS mixing within a single utterance was reported, which can result from the specific nature of the highly demanding NNS-NNS online interactions. Another notable finding is participants’ resorting to English through code-switching and through syntactic, lexical, and pragmatic transfer. The pedagogical implications concern the opportunities created by collaborative online intercultural interactions for authentic language practice and enhanced interactive reciprocal learning, particularly important in the context of a LCTL.

We look forward to seeing you there!