Errors in Coherence and Logic Embodied in the Microsoft Machine Translation Text— Based on the corpus of the Chinese-English translation texts taken from the 30th Han Suyin International Translation Contest

Based on the corpus of the original text, the Microsoft machine translation text, the human-translated text as well as the reference text taken from the 30th Han Suyin International Translation Contest, the paper establishes the error classification standard from three language levels and then focuses on the typical errors in coherence and logic embodied in the Microsoft machine translation text from the discourse level, aiming to provide the Microsoft machine translation system with reference and enlightenment for better translation quality and accuracy in the future. 微软机器翻译中行文连贯与逻辑思维问题研究 ——以第三十届韩素音国际翻译大赛汉译英文本为语料