Interpreting (i.e., oral translation) is facing challenges due to the prevalence of machine translation, and one urgent question for interpreting educators is what knowledge and skills should be taught in the machine-translation age. Focusing on this issue, the present corpus-based study systematically quantified the differences between interpreting outputs from expert interpreters and two machine translation systems. Using text analysis tools, significant differences in multidimensional linguistic features including lexical, syntactic, and cohesive ones were shown between humans and machines but not between two artificial systems. Finer-grained statistical analyses indicated that human-machine differences in certain indices deviated in varied interpreting modes. Our data collectively revealed the strengths of human interpreters in audience-oriented communicative mediation but limitations in cognitive resources. By relating the findings to interpreting competence, the current research provides important implications for empowering students in adaptively resorting to human strengths and/or embracing machine translation technologies.

中文翻译：

---

人机汉英口译输出的多维比较：对机器翻译时代口译教育的启示

由于机器翻译的盛行，口译（即口译）面临着挑战，口译教育者面临的一个紧迫问题是机器翻译时代应该教授哪些知识和技能。针对这个问题，本基于语料库的研究系统地量化了专家口译员和两个机器翻译系统的口译输出之间的差异。使用文本分析工具，人类和机器之间显示出多维语言特征（包括词汇、句法和衔接特征）的显着差异，但两个人工系统之间却没有显着差异。更细粒度的统计分析表明，不同解读模式下某些指标的人机差异存在偏差。我们的数据共同揭示了人类口译员在以受众为导向的沟通调解方面的优势，但在认知资源方面存在局限性。通过将研究结果与口译能力联系起来，当前的研究为赋予学生适应性地利用人类优势和/或采用机器翻译技术提供了重要的启示。