Perception of Japanese segmentals by native speakers of English

本研究试图通过研究英语本族语者如何感知日语语音来考察日语和英语元音和辅音的感知相似性。研究以英语为母语的人对日语的认知的研究远远少于英语母语者对日语的认知。本研究包括三个部分：对Mora鼻音和阻塞音的感知，对日语和美国英语元音相似性的感知，以及英语母语者对日语辅音的感知。四个以日语为母语的人产生了Mora obstrucent/Q/和Mora nasal/N/in/CVNCV/和/CVQCV/框架，其中/N/或/Q/后面的辅音是句号。最终的/CV/这个词被编辑掉了，产生了结构为/CVN/和/CVQ/的刺激。以英语和日语为母语的人听到了刺激，并确定了/N/和/Q/的发音位置。说英语的人比说日语的人表现得更好，这在很大程度上是因为在英语中，发声后鼻音和发声后停顿之间存在音位对比，而对于说日语的人来说，这些变化是音素差异。在第二部分中，两组受试者根据各自的母语元音类别识别对方语言中的元音。说英语的人为每个日语元音选择光谱上最接近的英语元音，但说日语的人也倾向于依赖持续时间信息。在另一个实验中，说日语的人为每个日语元音选择最接近的英语元音，他们倾向于将两个Mora/ii/和/uu/等同于/i/和/u/，将一个Mora/i/和/u/等同于/i/和/u/。在第三部分中，讲英语的人听日语单词和非单词，并将它们转录成英语字母表。他们的答案不一定与日语在英语中的常见抄写一致：例如，液体更多地被识别为/i/或/d/，而不是/r/。

This study attempts to investigate perceived similarity of Japanese and English vowels and consonants by focusing on how native speakers of English perceive Japanese speech sounds. There are far fewer studies examining the perception of Japanese by native speakers of English than the other way around. This study consists of three parts: perception of mora nasal and obstruent, perceived similarity of Japanese and American English vowels, and perception of Japanese consonants by native English speakers. Four native speakers of Japanese produced mora obstruent /Q/ and mora nasal /N/ in /CVNCV/ and /CVQCV/ frames where the consonant after /N/ or /Q/ is a stop. The word final /CV/ was edited out, creating stimuli with the structure of /CVN/ and /CVQ/ . Native English and Japanese speakers heard the stimuli and identified the place of articulation of /N/ and /Q/ . The English speakers outperformed the Japanese speakers largely because in English phonemic contrasts exist among postvocalic nasals and among postvocalic stops whereas for Japanese speakers these variations are allophonic. In the second part, the two groups of subjects identified vowels in each other's languages in terms of their own native vowel categories. The English speakers chose the spectrally closest English vowel for each Japanese vowel, but Japanese speakers tended to rely on durational information as well. In another experiment, the Japanese speakers chose the closest English vowel for each Japanese vowel, and they tended to equate two-mora /ii/ and /uu/ with /i/ and /u/, and one-mora /i/ and /u/ with /I/ and /U/. In the third part, the English speakers heard Japanese words and non-words and transcribed them in English alphabet. Their answers did not necessarily agree with the common transcription of Japanese in English: The liquid, for instance, was more often identified as /I/ or /d/ than /r/.