Speaker
Description
The rapid integration of artificial intelligence (AI) into language education has generated a growing body of research on tools, effectiveness, and pedagogical impact. Yet much of this work lacks explicit paradigmatic positioning, producing frequent misalignment between ontological assumptions, epistemological commitments, and methodological choices. The purpose of this paper is to propose a guiding framework that helps researchers identify and justify an appropriate paradigm when designing studies on AI in language education. Methodologically, the study adopts a conceptual, analytical approach: drawing on Cohen, Manion, and Morrison's Research Methods in Education, it reviews the relationships among ontology, epistemology, and methodology, and then systematically maps three dominant paradigms — positivist, interpretive, and critical — onto the distinctive features of AI as a research object, including its "black-box" opacity, the learner's role as an active agent, and questions of equity and ethics. The expected outcome is a three-tiered orienting matrix that links types of research questions to their corresponding paradigms, research designs, and data-collection methods, illustrated with concrete examples situated in language teaching and learning contexts. In terms of implications, the framework offers researchers, graduate students, and educators a practical tool for strengthening methodological transparency and coherence, reducing paradigm–method mismatch, and grounding their design decisions in a sound philosophical basis. By making paradigmatic reasoning explicit, the study contributes to more rigorous and reflexive inquiry at a time when AI is rapidly reshaping language education.
Keywords: research paradigms, methodology, AI in language education, ontology, epistemology
