An investigation of vowel devoicing and prosodic structure in Japanese

在这个项目中，我研究了元音脱音，这是一个研究得很好的日语音系特征，重点研究了它与韵律结构的相互作用，这是一个文献中相对未开发的领域。以下两个发现构成了我的项目的主要贡献。首先，我做了一个感知实验，测试说日语的人是否能听出有元音重音的单词之间的口音差异。我发现，尽管参与者总体上能很好地听出对比，但有一种对比类型比其他类型更难察觉，这表明他们对某些口音模式的感知存在偏见。其次，我进行了一项词典调查和生产实验，以发现是否像文献中声称的那样，元音的降元音会受到词内边界等形态学信息的影响。结果表明，影响日语元音消音的不是形态学，而是音系，特别是韵律信息。

In this project, I investigated vowel devoicing, a well-studied phonological feature of Japanese, focusing on its interaction with prosodic structure, a relatively unexplored area in the literature. The following two findings comprise the main contribution of my project.First, I conducted a perception experiment to test whether Japanese speakers can hear accent contrasts between words involving accents on devoiced vowels. I found that although participants could hear the contrasts fairly well overall, one contrast type was harder to detect than others, suggesting a bias in perception toward certain accent patterns.Second, I carried out a dictionary survey and production experiment to discover whether vowel devoicing would be affected by morphological information such as word-internal boundaries, as claimed in the literature. The results suggest that it is not the morphology but rather phonology, specifically prosodic information, that affects vowel devoicing in Japanese.