非對比特徵還是冗餘特徵的增強？

Non-contrastive Features or Enhancement by Redundant Features?

在漢語中，當詞綴 -r附加於以-n或-N結尾的音節時，-n和-N都會失落，但只有-N之前的元音會發生鼻化。近來有人提出（Zhang 2000）這是由於-r尚未附加時，-N的元音就比-n的元音鼻化程度大，並且這種原本非音位的差異在附加了詞綴以後變得有區別力。本文則主張，-r尾尚未附加時，帶-N的形式和帶-n的形式便已經具有音位上的區別，即軟顎鼻音的喉音化；此外，在以中、低元音收尾的所謂「開」音節中，其結尾處也有一個喉腔流音/A9/。這項主張可以經由與昌黎方言的比較而得到支持。在昌黎方言中，-n與-N兩個鼻音尾在附加-r之後都消失了，但二者的分別仍一致地保留下來。對於確認喉腔流音/A9/為一可能音位，以及在區別性特徵理論中，用舌根徵性 [±RTR] 取代舌體徵性 [±low] 的更進一步的證據來自漢語的歷史以及其他語言（包括英語）。在英語中，[±RTR] 取代了傳統鬆／緊徵性的對立來描述長短元音的差異。這證實了根據標準的最小對比測試，特徵的對立嚴格說起來應為冗餘特徵的概念，而冗餘特徵在容易造成混淆的環境中，具有增強區別力的作用。

In Mandarin Chinese, when the suffix -r is added to syllables ending in -n and -N, both consonants are dropped but the vowel is nasalized in the case of N. It has recently been argued (Zhang 2000) that this is because in the unsuffixed forms there is already a greater degree of nasalization in the case of -N than in the case of -n and that this merely allophonic contrast becomes distinctive when the suffix is added. It is here claimed that there is already a phonemic distinction present in the unsuffixed forms, namely pharyngealization of the velar nasal, and that furthermore there is a pharyngeal glide /A9/ coda in the case of so-called ‘open’ syllables ending mid and low vowels. This is supported by comparison with the Changli dialect in which both nasal finals are deleted before -r but the distinctions are uniformly preserved. Further evidence for the need to recognize the pharyngeal glide /A9/ as a possible phoneme and for replacing the dorsal feature [±low] with the tongue-root feature [±RTR] in distinctive feature theory is adduced from the history of Chinese as well as from other languages, including English, where it takes the place of the traditional tense/lax contrast in short and long vowels. This supports the concept of featural contrasts that are strictly speaking redundant according to the standard test of minimal pairs but that serve to enhance distinctions that would otherwise be easy to confuse.