Numerous studies demonstrated that phonological short-term memory (PSTM), measured via nonword repetition tasks, strongly predicted vocabulary learning (Service, 2012). Complex WM span tasks (e.g., reading span, operation span) predicted higher-order syntactic processing and sentence comprehension (Harrington & Sawyer, 2001). Critically, research showed that WM and traditional aptitude tests (MLAT) overlapped but were not identical. Linck et al. (2014) conducted a meta-analysis confirming that WM explains unique variance in L2 outcomes beyond the MLAT, particularly in the early stages of acquisition.
This research effectively expanded the aptitude construct. Aptitude was no longer just “learning ability” but included the online cognitive machinery necessary for real-time language processing. 3. Aptitude-Treatment Interactions (ATIs): Matching Learner to Method (2010–2018) If aptitude is multidimensional, then different learners should thrive under different instructional conditions. This led to a resurgence of Aptitude-Treatment Interaction (ATI) research. The classic hypothesis—that high-analytic learners benefit from explicit grammar instruction while high-memory learners benefit from immersion—was refined. twenty-five years of research on foreign language aptitude
The past twenty-five years have witnessed a remarkable renaissance. Researchers have moved beyond simple prediction to ask deeper questions: How does aptitude interact with instructional conditions? Is aptitude a unitary construct or a constellation of flexible resources? Can it be developed? This paper synthesizes the key empirical and theoretical contributions to FLA research from 1999 to 2024, organizing the literature into four thematic waves. The first major shift was the integration of working memory (WM) into the aptitude framework. While traditional aptitude tests emphasized crystallized knowledge and analytical reasoning, WM—the ability to simultaneously store and process information—offered a process-oriented explanation for individual differences. Linck et al
Using idiodynamic methods (moment-to-moment ratings), Suzuki (2021) showed that learners’ effective WM capacity fluctuates depending on perceived task difficulty and state anxiety. A learner who appears “low aptitude” on a timed grammaticality judgment test may perform as “high aptitude” on a self-paced narrative retell task. Aptitude was no longer just “learning ability” but