Article Preview
TopLiterature Review
The evolution of NMT has sparked a paradigm shift in both the practice and theoretical discourse of translation, particularly concerning the efficiency and cognitive demands of PE strategies. Pioneering studies, such as those by Koponen (2012) and O’Brien (2006), highlight the pivotal role of NMT in elevating machine translation quality, thereby transforming traditional translation workflows and emphasizing the indispensability of PE. Nevertheless, the efficacy of the PE process depends on the quality of machine translation outputs and the translator’s expertise. Subsequent investigations have scrutinized the influence of generative artificial intelligence platforms, such as ChatGPT, on machine translation paradigms. Despite notable advancements in specific domains, such as Chinese-to-English translation, terminological precision, and literary transcreation, GPT technologies have not surpassed traditional NMT frameworks completely (Zhang & Zhao, 2024). In a fresh approach, Wei and Chen (2023) introduced an entropy-weighted Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method for assessing machine translation quality, calling for a multidimensional evaluation framework.
Amid the enhanced translational throughput enabled by NMT, translators continue to tackle the intricacies of interpreting machine-generated content. The scholarly focus has progressively shifted toward the threefold goal of optimizing translation quality, minimizing time expenditure, and reducing cognitive load. Landwehr, et al.’s (2023) research on Open TIPE not only highlighted the potential of interactive PE environments to bridge human-machine collaboration gaps but also underlined the significance of harnessing human insights to refine APE models, fundamentally altering PE dynamics. Bundgaard’s (2017) investigation into translator attitudes toward TCI emphasized the nuanced responses to MT-assisted TM systems, revealing a blend of skepticism and pragmatic adaptation among professionals. Kenny & Doherty (2014) praised the benefits of statistical machine translation in boosting PE efficiency, but concurrently recognized its shortcomings in ensuring consistency and precision. Furthermore, Garcia’s studies (Garcia, 2010a, 2010b, 2011a, 2011b) outlined the evolving competencies and role dynamics of translators in the NMT era. A persistent challenge is the balancing of machine translation’s efficiency with the superior quality inherent in human translation, particularly when addressing complex or domain-specific texts. This dilemma extends to how translators can wisely allocate cognitive resources during PE to enhance translation efficiency while alleviating cognitive strain.
Current scholarship reveals a notable gap in quantitatively outlining the correlation between translators’ cognitive load and translational efficiency. Therefore, in this study, I combine qualitative and quantitative methodologies to thoroughly explore decision-making processes, time management, and error rectification strategies in PE. I focus particularly on the cognitive tactics employed by translators in the context of NMT, analyzing their repercussions on translation quality and efficiency. The distinctiveness of this study lies in its comprehensive consideration of not only PE efficiency but also the cognitive load carried by translators. By conducting empirical analyses, I clarify the cognitive dynamics and strategic preferences of translators in the PE phase, highlighting their impact on both the quality and efficiency of translation. Consequently, I aim to provide novel theoretical insights and practical recommendations for advancing machine translation technology and translation education.