音位分析的“多解论”和最佳答案

The Non-uniqueness Theory and the Best Solution in Phonemic Analysis

音位是语言学的一个基本概念,但一个语言有哪些音位往往存在争议。Chao(1934)提出,音位分析有“非唯一性”,即有“多解”。比如,Lee&Zee(2003)认为汉语普通话有21个元音,而Duanmu(2007:35)认为只有5个(不包括儿化音)。对“多解论”来说,两种分析都可以,无优劣之分。“多解论”从理论上放弃了对最佳答案的追求,于是人们往往众说纷纭。因此,又有人开始对音位分析产生质疑,甚至提出放弃音位这个概念(Ladefoged2001:170;Fowler2015)。“多解论”有个误区,即将音位分析与其他音系现象剥离开来,如押韵现象、韵母结构、音节空挡等。如果将音位分析和其他音系现象联系起来,那么音位分析没有多解性,最佳答案是相当清楚的。文章以汉语普通话为例,全面比较各种分析,以论证以上的结论。

The phoneme is a basic concept in linguistics, yet opinions often differ over what phonemes a language has. Chao (1934) proposes that, because there are often competing criteria, there is no unique solution in phonemic analysis (hence the ‘non-uniqueness theory’). For example, Lee & Zee (2003) propose that Standard Chinese (Putonghua) has 21 vowels, whereas Duanmu (2007:35) proposes that Standard Chinese has just five vowels. From Chao’s perspective, both proposals are valid, and neither is better than the other. Chao’s proposal offers a theoretical justification for giving up the search for the best solution in phonemic analysis. As a result, various phonemic analyses of the same language often co-exist, with no consensus in sight. This in turn has led some linguists to doubt the validity of phonemic analysis, and some even propose to give up the concept of the “phoneme” altogether (e.g. Ladefoged 2001, Fowler 2015). I argue that there is a flaw in the non-uniqueness view, in that it fails to place phonemic analysis in the context of other patterns of phonology, such as riming patterns, the structure of the syllable rime, and syllable gaps. If we consider phonemic analysis in relation to other phonological properties, then there is no non-uniqueness, and the best solution is not only obtainable but quite clear. For illustratation, I shall use Standard Chinese as an example and evaluate various analyses, in order to determine the best solution.