The study examines the relationship between working memory (WM), executive control, and conference interpreting performance across consecutive interpreting (CI) and multiple forms of simultaneous interpreting (SI). Using a nine-month longitudinal design, the study compares interpreting training alone with interpreting training combined with targeted WM training, tracking changes in both cognitive resources and interpreting outcomes over time. Performance was assessed in CI, classic SI, SI with Text, and SI with PowerPoint (PPT), alongside a multidimensional battery measuring WM, executive control, and verbal fluency.
Across the interpreting battery, WM emerged as the most consistent predictor of performance, whereas executive control and verbal fluency contributed in more context-dependent ways, varying with interpreting mode, directionality, and support configuration. Visual support did not function as a uniform aid. Text-based support facilitated SI under conditions of high speech–text alignment, whereas PPT support yielded mixed effects, reflecting a trade-off between cueing benefits and increased coordination and inhibition demands. Directionality patterns further indicate that L2-target interpreting places greater demands on limited WM and executive-control resources than L1-target interpreting.
To account for these findings, the thesis proposes a Unified Cognitive Process Model of Conference Interpreting that integrates WM components, executive control, and modality-specific coordination demands within a single architecture. The model frames visual support effects as configuration-dependent reallocations of processing demands rather than as uniformly load-reducing.
Training findings indicate that interpreting practice strengthens verbal maintenance and language-embedded skills, whereas targeted WM activities preferentially enhance domain-general control processes such as updating and attentional regulation. Task-based dissociations between computerised executive-control measures and verbal fluency further highlight the need to distinguish domain-general and language-embedded cognitive skills when designing WM training for interpreters. Overall, the study advances a configuration-based account of multimodal support in interpreting and offers implications for cognitive theory, interpreter training design, and future research on time-pressured bilingual processing.
