Rethinking Second Language Anxiety: Looking Beyond the Classic View of the Anxiety → Achievement Relationship

A large and emerging body of evidence has challenged the classic view that L2 anxiety is a causal factor in L2 learning achievement and that when anxiety is mitigated, L2 proficiency improves. This classic “anxiety → achievement” view has three major problems. First, studies frequently draw causal conclusions from cross‐sectional associations, even though such designs cannot establish temporal precedence and are highly vulnerable to reverse causality and confounding. Second, common L2 anxiety self‐report instruments embed language skills and self‐perceived competence within item content, inflating anxiety–achievement correlations. Third, the classic view on anxiety has developed an “anxiety monoculture”: a persistent tendency to favor the affective interpretation “anxiety → achievement” while neglecting rival explanations, language‐related confounds, and falsification‐oriented designs. In this paper, we synthesize evidence from longitudinal and experimental work indicating that language achievement more plausibly precedes anxiety than the reverse, and provide pedagogical implications that highlight emotional support as an important resource in the process rather than a key causal factor. We then propose an alternative research agenda for the new view on L2 anxiety comprising four components: (a) stronger causal‐inference designs, including longitudinal and experimental approaches; (b) improved measurement standards that decouple anxiety from language ability; (c) explicit accounting for language achievement and L2 aptitude measures when modeling the anxiety–achievement relationship; and (d) reconceptualizing anxiety as a reflection of the learner's concurrent level of language proficiency, which can decrease as language competence and language achievement grow. We argue that progress in conceptualizing L2 anxiety depends crucially on moving beyond past designs and interpretations by adopting these recommendations.