This study investigated the use of Artificial Intelligence (AI) models and signal detection processes to generate meaningful visual and ChatGPT‐like narrative feedback on second language (L2) English intelligibility. To test the effects and perceptions of such techniques, three groups of learners (N = 90) received visual and narrative feedback (n = 30), visual‐only feedback (n = 29), and no feedback (n = 31) in an online self‐paced intervention with explicit instruction on segmental and suprasegmental features of intelligibility. Pre/postspeaking tasks were evaluated by raters for intelligibility, comprehensibility, and accentedness, as well as segmental and suprasegmental accuracy, in scripted and spontaneous speech. The results indicate that visual feedback improves prominence production, but only those participants who also received the narrative (i.e., ChatGPT) feedback improved in two of the three prosodic features and in intelligibility. However, those who received narrative feedback had the lowest perceptions of the practice activity helpfulness. Implications for the use and improvement of AI‐based pronunciation feedback are provided.
This study proposed and tested a revised model of feedback-seeking behavior in second language acquisition. The revised model included feedback elicitation, monitoring, and use. In two studies, data were collected from 207 learners enrolled in Spanish classes at a large public university in Florida and 294 learners of English at a multi-branch private language institute in Iran. Exploratory factor analyses generally combined CF Monitoring and CF Use, whereas confirmatory factor analyses favored the hypothesized three-factor model, with stronger support in Study 2. The scales demonstrated acceptable convergent validity and internal consistency, although the distinction between Monitoring and Use was limited in Study 1 and improved in Study 2. Hierarchical multiple regression analyses further showed that the overall FSB index significantly predicted self-assessed L2 proficiency in both contexts, although the unique contributions of the individual components varied across samples. Overall, the findings supported the tripartite model of FSB, defined as learners’ proactive efforts to elicit, monitor, and use CF to enhance their L2 production. Theoretical and practical implications are discussed, and future research directions are outlined.
This paper reports on the development of academic-specific L2 target benchmarks and introduces a web-based instantaneous assessment tool to provide second language (L2) English learners with quantitative results of their utterance speech rate and pauses. Thirty highly intelligible L2 speakers of English were recorded reading a paragraph aloud and responding to an explanation-based unconstrained speaking prompt to provide benchmarks for learners. Then, six pre-service International Teaching Assistants (ITAs) completed a Moodle course on pause units and speech rate with practice tasks that use the embedded instantaneous assessment tool. Speech captured at a pre-and post-tests was analyzed acoustically and evaluated by six trained human raters for appropriate speech rate, pause units, accentedness, and intelligibility. The results indicate that participants were able to apply the practice to their use of pause units and converge on the target speech rate, and that the changes resulted in trends of increased intelligibility but not accentedness. The paper gives insights into employing automated fluency analysis tools and digital pedagogical content for L2 learners, offers additional evidence of the relationship between suprasegmental features and listener perceptions, and provides support for the use of instantaneous assessment tools with L2 learners.