Examining Cognitive Load in Human-Machine Collaborative Translation: Insights from Eye-Tracking Experiments of Chinese-English Translation

Lei Chen^*

• Jinjiang College, Sichuan University, Meishan, China

With the development of artificial intelligence and computer science, human-computer collaborative translation (HMCT) mode has gradually become a research hotspot in the field of English translation. The purpose of this study was to explore the cognitive load characteristics of translators in the process of human-computer collaborative translation through eye tracking experiments of Chinese-English translation. Based on a 2×2 hybrid design, the participants' eye movements were analyzed under the conditions of simple, medium and complex texts through two tasks, human translation and human-computer collaborative translation. The study involved 30 master's students or translators in translation who used Tobii Pro Glasses2 to record eye tracking data in real time, focusing on fixation time, regressionness, saccade and fixation point to reveal the impact of different Chinese-English translation tasks and text types on cognitive load. The experimental results show that the fixation time, the numbers of regressions, fixations and saccades of human translation are significantly higher than those of human-computer collaborative translation, especially in complex text tasks. At the same time, the numbers of regressions and fixation time increased significantly with the increase of task complexity in both groups, and the human translation group showed a higher cognitive load in complex tasks. This study finds that the cognitive load of translators in the process of human-machine collaborative translation shows phased changes, especially when the output quality of machine translation is poor, translators need more cognitive resources to correct. The impact of complex tasks on cognitive load is even more significant, and human translation requires more cognitive effort on the part of translators. Eye tracking data analysis provides empirical support for understanding the cognitive mechanisms in the translation process. For the first time, this study systematically explored the cognitive load characteristics of human-computer collaborative translation through eye tracking technology, filling the research gap in this field in the existing literature. The results of this study not only provide a theoretical basis for optimizing translation tools and designing more efficient translation processes, but also provide a new perspective for cognitive load management in translation teaching and practice.