Exploring Cognitive Processing in Simultaneous Interpreting with Live Captioning: A Pilot Study on Ear-Voice Span and Output Accuracy

Abstract Simultaneous interpreting (SI) with live captioning (LC) is an emerging working mode that has gained traction among both practitioners and researchers. While previous studies have examined LC’s impact on output accuracy and physiological responses of interpreters, ear-voice span (EVS) — the time lag between the speaker’s utterance and the interpreter’s production — remains underexplored, despite its dual role as a window into interpreters’ cognitive processes and a tool for managing cognitive resources. This study investigates LC’s influence on EVS and output accuracy, providing insights into cognitive mechanisms through the lens of Gile’s Effort Model.A trial study was conducted with two professional interpreters — one frequent and one infrequent LC user — who performed SI with and without LC, followed by a semistructured interview. The results pointed to the following findings. 1. LC affected EVS and output accuracy differently between the two interpreters: the frequent LC user exhibited a slightly shorter EVS and marginal accuracy improvement, while the infrequent LC user experienced a significantly longer EVS and reduced accuracy. 2. LC serves as an aid to interpreters, but performance remains determined by core SI skills, such as active listening, analysing, deverbalizing and reformulating. 3. LC is most effective as a support for the memory effort but can be detrimental if used as a substitute for listening. 4. Both excessively short and long EVS indicate that the interpreters’ normal listening and chunking are disrupted by an overreliance on reading captions.The findings underscore the need for targeted training to help interpreters develop strategies, including sound regulation of EVS, for balancing listening and reading captions in this new working mode. The study provides both theoretical and practical implications for interpreter training and the integration of AI-powered tools into professional SI workflows.