Deep Learning–Oriented Speaking Instruction: The Role of Memory, Cognitive, and Compensation Strategies in EFL

Abstract

This study examines memory, cognitive, and compensation strategies in deep learning–oriented English-speaking instruction within an EFL context. Grounded in Oxford’s direct language learning strategy framework and deep learning pedagogy, speaking is conceptualized as a cognitively and strategically regulated process. Memory strategies facilitate encoding and retrieval, cognitive strategies support the processing and organization of ideas, and compensation strategies enable learners to sustain communication under constraints. The study aims to analyze how these strategies are enacted and interact in learners’ speech production and to explain their contributions to fostering deeper, more reflective, and more autonomous speaking practices. Adopting a qualitative descriptive design, data were collected through classroom observations and semi-structured interviews involving 34 English Literature students at Universitas Negeri Makassar (UNM), cohort 2024, Class A. The findings indicate that memory strategies support lexical retrieval and fluency, cognitive strategies facilitate idea development and discourse organization, and compensation strategies help learners manage communicative pressure and maintain interactional flow. Importantly, the interaction among these strategies promotes deeper engagement in speaking tasks, reflecting key principles of deep learning. The study suggests that deep learning–oriented speaking instruction, aligned with Kurikulum Merdeka, should explicitly integrate strategy-based pedagogy to support reflective, autonomous, and meaningful oral communication. These findings contribute to EFL pedagogy by demonstrating how language learning strategies function collectively to support sustainable speaking development in higher education contexts.