Pixel International Conferences

In Italian higher education, the teaching of Pragmatics remains underdeveloped and often addressed through traditional, form-focused approaches [1]. The need to expand language practice and pragmatic awareness, while enhancing university students' engagement, reducing in-class time and providing learners with personal tutors, has led to the development of LLM-based chatbots, currently representing one of the most promising tools within Technology-mediated Second Language [2] learning. Task-Based Language Teaching [3] approach, within the CALL [4] framework, can be leveraged to design chatbots able to enhance the development of pragmatic competence in Italian L2 task-based scenarios. Chatbots may allow structuring task-based service-encounter interactions and act as conversational partners that provide immediate, context-sensitive feedback, aligned with the CEFR descriptors [5]. We are currently investigating two alternative ways of implementing a conversational tutor: Custom GPT/Gem or Agentic Application. The former is created under the name “Pragmabot”, and is currently being evaluated. Pragmatics, however, is highly context-dependent and socially situated, which makes it particularly difficult to shape in black-box custom conversational agents, typically more effective at handling linguistic form than pragmatic use. The latter implementation option consists in a more controllable tool currently under design: the agentic application PragmAI would be guided by an explicit representation of pragmatic knowledge, and refer to a database of learners’ characteristics, needs and difficulties, to foster a personalized interaction for university students in the aforementioned communicative context.

Keywords: Pragmatics; Chatbot; Oral interaction; TBLL; CALL; Technology-mediated Second/Foreign Language (L2) Pragmatics.