A survey of twelve translation students in 2017 revealed that they tend to find translating figurative and metaphorical language difficult. In addition, an experiment also conducted in 2017 showed similar results. During the first phase of this experiment, two trained researchers coded metaphorical items in a text from the New Scientist following the Metaphor Identification Procedure Vrije Universiteit Amsterdam (MIPVU). Based on Cohen’s kappa, the researchers reached an initial coding agreement of 0.692 (strong agreement) and a final agreement score of 0.958 (almost perfect agreement) after discussion. The second phase of the experiment involved the coding of the metaphorical items previously identified by the researchers in the same text by 47 students who received a two-hour introduction to conceptual metaphor theory and a simplified method to code metaphorical items. However, the results of the students’ coding showed that they had failed to identify metaphors in 49.96% of cases. Nevertheless, a chi-squared test (p < 2.2-16) revealed that the students’ coding was not due to chance alone and therefore not arbitrary.

中文翻译：

---

语言隐喻的识别和翻译困难：翻译学生的调查和实验

2017年对12名翻译学生的调查显示，他们往往觉得很难翻译比喻和隐喻语言。此外，2017年进行的一项实验也显示了相似的结果。在该实验的第一阶段，两名受过训练的研究人员按照阿姆斯特丹隐喻识别程序（MIPVU），在《新科学家》的文本中对隐喻项目进行编码。经过讨论，研究人员根据科恩的kappa得出初始编码协议为0.692（强协议）和最终协议分数为0.958（几乎完美的协议）。实验的第二阶段涉及对47名学生在同一课文中先前由研究人员确定的隐喻项目进行编码，他们接受了两个小时的概念隐喻理论介绍和简化的隐喻项目编码方法。但是，学生编码的结果表明，他们在49.96％的案例中未能识别隐喻。然而，卡方检验（p <2.2-16）表明，学生的编码并不是仅仅由于偶然性，因此也不是任意的。