Feedback Valence and Framing in AI-Mediated EFL Learning: A Quantum-Inspired Analysis of Their Effects on Goal Orientation, Motivational Affect, and Task Persistence Through Achievement Goal Theory

Abstract

As artificial intelligence (AI) increasingly shapes language learning, the role of feedback in AImediated environments has become a focal concern-particularly in how it influences learner motivation, affect, and engagement. Grounded in Achievement Goal Theory and enriched by a quantum-inspired analysis, this mixed-methods study examined how feedback valence (positive vs. negative) and framing (process-oriented vs. outcome-oriented) jointly impact English as a Foreign Language (EFL) learners' goal orientation, motivational affect, and task persistence. The study also explored non-classical cognitive patterns, including emotional ambivalence, decision reversals, and motivational interference. A total of 120 undergraduate EFL students were randomly assigned to one of four feedback conditions during an eight-session ChatGPT-based grammar course. Quantitative data were gathered through validated instruments measuring goal orientation, motivational affect, and task persistence. Qualitative data from reflection logs and semi-structured interviews were analyzed thematically to uncover deeper cognitive-emotional dynamics. Results from MANOVA and follow-up ANOVAs revealed that positive, processoriented feedback significantly enhanced mastery goals, positive affect, and persistence, whereas negative, outcome-oriented feedback resulted in declines across these domains. Qualitative findings uncovered complex, nonlinear responses including dual emotional states, motivational conflicts, and cognitive interference, which are thepatterns consistent with quantuminspired models of cognition. This study offers both theoretical and practical implications, highlighting the importance of feedback design in AI-supported instruction. It underscores how subtle variations in feedback framing and tone can generate divergent motivational trajectories, and introduces a novel quantum-inspired lens to capture the probabilistic, emotionally dynamic nature of learner cognition in digitally mediated settings.