【摘要】 藏语拉萨话的声调如何划分一直是学界讨论的热点话题，主要分歧在于T53和T51、T23和T232是否应该合并。本文通过感知实验发现，无论基频如何变化，母语者都无法区分声调对立组内的两个声调。因此，对于有声调和喉塞韵尾两项对立的音节而言，声调是羡余性特征，喉塞韵尾是区别性特征。本文认为可以将T51和T53归为一个调位，将T23和T232归为一个调位，藏语拉萨话的声调应该划分为四个调类。通过对藏语拉萨话声调的感知研究可以看出，现代语音学和认知音位学的研究方法不仅可以解决传统结构主义音位学无法确定音位的问题，也可以解决不同调查者调查同一语言时可能出现音位系统差异的问题。更多还原

【Abstract】 The classification of Lhasa Tibetan tones has been a hot topic in the academic discussion, with the main disagreement being whether the two sets of tones, T53 and T51 as well as T23 and T232, are four separate tones or just two distinct tones with allotones. Through perceptual experiments, we find that no matter how the fundamental frequency changes, native speakers cannot distinguish the two members in each tonal sets. Therefore, for morphemes involving both tonal distinction and presence or absence of the final glottal stop, tonal contrast is a redundant feature while the presence or absence of the final glottal stop is the distinctive feature. We propose that T51 and T53 are allotones of one toneme, while T23 and T232 are allotones of another; hence the six tones can be classified into four tonal categories. The perceptual study on Lhasa Tibetan tones shows that the research methods of modern phonetics and cognitive phonemics can not only solve the dilemmas of determining phonemes in traditional structuralist phonemics, but also resolve differences in phonemic systems obtained by different researchers investigating the same language.更多还原