音位、特徵和音節：如何將聲韻母轉為元輔音

Phonemes, Features, and Syllables: Converting Onset and Rime Inventories to Consonants and Vowels

John A. Goldsmith認為，到目前為止，音系學的最大成就仍然是音位分析，但是下一步我們應該將注意力轉向小於音位的「特徵」和大於音位的「音節」。我們認為特徵和音節的研究不但重要，而且也是音位分析的基礎。音位分析中的很多疑難問題，如果沒有特徵理論和音節理論，就很難解決。我們以蘭州話的分析為例，指出沒有音節和特徵的概念，音位分析往往有多種答案，而且很難對其進行選擇，正如趙元任所說。如果將音節和特徵帶入音位分析，不僅能夠大大減少可行的答案數量，而且可以探討一系列新的問題，特別是解釋音節的空檔問題（即哪些音節會出現，哪些音節不會出現）。

Phonemic analysis has been thought to be the 'greatest achievement' in phonology; in contrast, progress at other levels, such as the feature level below and the syllable level above, has been limited. We argue that not only are features and syllables important, but phonemic analysis is inadequate without reference to them. In other words, features and syllables are not separate areas of inquiry, but solve fundamental problems in phonemic analysis. We demonstrate the point with an in-depth analysis of Lanzhou Chinese. We show that, without reference to features and syllables, phonemic analysis is ambiguous in that it is open to alternative solutions, a problem noted by Yuen Ren Chao. In addition, there is no explanation for the patterns of occurring and non-occurring syllables. In contrast, if we take features and syllables into consideration, not only can we account for occurring and non-occurring syllables, but the phonemic analysis itself becomes simpler and less ambiguous.