Research on Mining Development Patterns of Pragmatic Competence in English Learners Assisted by Machine Learning

Abstract

With the growing demand for cross-cultural communication in the process of globalization, the cultivation of pragmatic competence in English learners has become a critical component of language education. Traditional pragmatics research, which predominantly relies on manual annotation and small-sample analysis, struggles to systematically reveal the developmental trajectories of pragmatic competence. This study, situated at the intersection of pragmatics theory and computational linguistics, conducts research on mining development patterns of pragmatic competence in English learners assisted by machine learning. By constructing a theoretical framework encompassing three dimensions—linguistic pragmatics, social pragmatics, and cognitive processing—the research systematically analyzes explicit and implicit pragmatic features. Subsequently, techniques such as fine-tuning pre-trained language models, multi-modal fusion, and graph neural networks are employed to achieve automated extraction and pattern recognition of pragmatic features. Finally, through multi-dimensional validation and empirical teaching research, the scientific validity and application value of the identified developmental patterns are confirmed. This study offers a new perspective for understanding the mechanisms of pragmatic competence development and provides methodological support for achieving precision in pragmatic instruction.