The Dynamic Nature of Emotions in Second Language Learning

Abstract

The Dynamic Nature of Emotions in Second Language Learning: Theory, Measurement, Prediction, and Intervention examines how emotions develop and fluctuate during second-language acquisition (SLA), challenging the traditional view of emotions as relatively stable learner characteristics. The dissertation argues that emotions should instead be understood as temporally structured, dynamic processes operating at multiple timescales: micro-level changes within individual communicative events and macro-level trajectories across days or weeks of learning. The theoretical chapter develops this dynamic framework and proposes intensive methods such as idiodynamic measurement, ecological momentary assessment (EMA), and time-series modelling. The first empirical study demonstrates that speaking anxiety occurs in brief, recurring episodes. Although speaking a second language produces stronger and more sustained anxiety and physiological arousal than speaking a native language, the underlying psychophysiological mechanism appears largely shared across the two contexts. A 21-day EMA study further shows that emotional experiences are more variable within individuals than between individuals. Anxiety, enjoyment, and boredom repeatedly form distinct emotional profiles, with “Routine but Pleasant” emerging as the most common everyday learning state. Importantly, recent emotional history allows relatively accurate short-term prediction of subsequent emotions, suggesting that emotional fluctuations are structured rather than random. Finally, a 20-day breathwork intervention demonstrates that brief cyclic sighing can reduce anxiety particularly around learning moments and may buffer the accumulation of negative emotions. Overall, the dissertation shifts SLA emotion research from static snapshots toward understanding when emotions arise, how they develop, how they can be predicted, and when interventions may be most effective.