Volume 24, No 1
Hung-nin Samuel Cheung 张洪年
For many centuries, our knowledge of ancient Chinese phonology has relied primarily on the study of rhyming patterns in particular poetic corpuses. The effort is based upon the presumption that poets and lyricists share not only the same language but also a common phonological awareness that allows them to choose rhyme words with the same or similar yunmu in their compositions. Any differences in practice may be construed as indicative of dialectal variations or of new developments in language. This paper challenges that view by examining the rhyming practice in close to five hundred popular Cantonese songs. The phonological system as reconstructed on the bases of thousands of rhyme words yields twenty-four finals, nineteen short of what we find in the actual spoken language. The results are alarming. Close analysis reveals that while phonological identity remains a strong preference in rhyming, it is not the precluding factor. Many words rhyme because they share the same vowel even though they may differ in their consonantal endings. Others interact for historical reasons and do not reflect any phonological changes in modern pronunciations. Cross rhyming allows literary flexibility but can be misleading in terms of what it informs us about the language. The paper also discusses the use of bilingual rhyming in lyrics that contain English words, a phenomenon that bespeaks the hybrid nature of speech in contemporary Cantonese.
研究古韵分部大都根据古人诗歌押韵用字，归纳类推而得。材料中用韵分合有异，往往会视作音韵变化的痕迹。然而古人用韵是否一定从 严？两字相押， 是否一定韵母相同？ 本文试就香港粤语歌曲中的押韵情形，对这问题作一探讨。研究材料包括四百多首流行歌曲，押韵的字数以千计。韵脚可以利用系联的原则归纳成二十四韵部，而现代粤语共有五十三个韵母，两者之间相距甚大。细究其因，粤语歌曲除韵母相同的字可以互相押韵以外，更有两种通押的现象。一种是凡元音相同的字皆可互押，所以阳声韵本-m，-n， -ng 三分，诗歌中可以通叶， 而且阴阳入 通押的例子也比比皆是。另一种情形是历史遗留下来的押韵成规。江阳不分，支微通叶，这是从词曲以来已有的现象，今人填词，虽是发音不同，但仍照叶无碍。据此而言，诗歌用韵和实际语言并不完全符合。研究古音，对于材料运用，不得不严加谨慎。香港粤语歌曲中又有中西语言互相押韵的现象，文中也一并陈述。
The Modern Shanghai Dialect has undergone drastic sound changes since the 1930’s. This paper proposes that the changes are results of contacts with immigrant and prestigious dialects that have influenced the local phonological system. It also provides an analysis of the characteristics of these changes.
In the study of pre-modern Chinese phonology, no single source has been more important than the Qieyun (QY) 切韵; and a correct understanding of the origin and nature of this text has consequently been a primary concern in the field. Perhaps the most significant modern study of these text-historical questions is that of Zhou Zumo 周祖谟 (1966). And this work in turn is best known to Western scholars through the translation of Malmqvist (1968). Indeed, the influence of Zhou’s original article, as represented in Malmqvist’s translation, has been of paramount importance in the West.
Judith N. Rabinovitch
This article introduces the reader to a now defunct Sino-Japanese language peculiar to Japan known as hentai kambun 变体汉文, “variant Chinese.” It was widely used along with orthodox Chinese and various literary Japanese styles in the Japanese court throughout the premodern period. Some two-thirds of all extant premodern Japanese texts, including government records and communiques, biographies, temples histories, and a range of literary materials, are written in some form of hentai kambun. A knowledge of it is therefore a sine qua non for the study of Japanese civilization. The linguistic characteristics of variant Chinese–its hybrid Sino-Japanese grammar and syntax, vocabulary and orthographical features–are described and illustrated with examples, and the origins and historical place of variant Chinese relative to literary Japanese and orthodox Chinese are also discussed. Fully naturalized styles of hentai kambun, often referred to collectively as kiroku-tai 记录体, “the documentary form,” emerged in the Heian period (ca. 9-12th c.) and were seen in their day not as sub-standard or aberrant forms of Chinese but instead as heavily Sinicized styles of written Japanese. The complex interplay between these kiroku-tai forms, kundoku 训读 (a system for translating Chinese texts into Japanese), and literary Japanese, both medieval and modern, is also treated here, with examples of various Chinese-derived locutions that are found in modern written Japanese. Finally, the article explores the important literary uses of variant Chinese in such genres as courtier diaries and war tales, as well as in Japanese poetic criticism. Variant Chinese literature is one of the least developed areas in Japanese studies. Our future progress will depend upon our ability to train new scholars in Japan and the West to use hentai kambun materials skillfully in their literary research.
In this paper we argue that the name of the temple, Bai Ma Si, 白马寺, comes from the Sanskrit word padma, which means ‘lotus’. The white horse was all but unknown as a symbol in ancient India, from whence Buddhism came to China; but it was a potent topic in Chinese thought at least since the famous philosophical discussion of Gongsun Longzi. To support this argument we note that the endings -t and -k were already merging or varying at least as early as Shijing times, thereby explaining why the pad- in Sanskrit was transcribed with the Old Chinese syllable brak, which eventually evolved into Putonghua bai. We further speculate that the -d sound in pad- indicated a low pitch first syllable in the Sanskrit word. The fact this syllable was transcribed with a b- in Chinese, rather than directly with a p-, shows the greater saliency of pitch over the segmental feature of voicing.
Y. R. Chao (1968) classified Chinese verbs into two major classes according to their transitivity and then each is broken down into further classes in accordance with such criteria as action, quality or state and so forth. With the advent of the case and its application to the study of Chinese syntax, researchers such as Li (1971), Tang (1975), Teng (1975) have all attempted at verb classification, taking the deep case information into consideration. Li’s study, which will be taken as representative here, reverts Chao’s system by taking the distinction of action vs. state/quality as the most important criterion, to be followed by the distinction between ‘transitive’ and ‘intransitive’. More recently, with more and more information available on verb subcategorization, both CKIP group of Academia Sinica (1988) and Her (1990) have made new attempts at verb classification, using such models as the case theory and Lexical Functional Grammar or some kinds of combination of the two. In this paper, we will examine all four classifications, i.e., Chao’s, Li’s, CKIP’s and Her’s, critically, pointing out their significant generalizations as well as inadequacies as we go along. In the end, on the basis of such criteria as the number of arguments, the type of arguments, theta roles, and special distributional restrictions, a classificatory system of 14 major classes will be proposed. The major features of each class will be explained with illustrative examples.
Volume 24, No 2
Qieyun system (QYS) palatal initials which are suspected of an Old Chinese (OC) velar origin are of two types: (1) Type I palatals occur in certain syllables with front vowels which are subject to the chongniu phenomenon: palatalized OC velars are in complementary distribution with chongniu division (“grade”) 4 syllables. Therefore, such palatals can be reconstructed as ordinary velars in OC, followed by whatever gave rise to QYS div. 4 chongniu medial and/or vocalism, e.g. 支 Zhī. < OC *ke. (2) Type II is the QYS initial tśhj which goes back to some initial cluster involving a velar and *1, with any vowel, e.g. 川 chuān < OC *k’lun(?).
中古切韵系统中有一组腭化音(palatal) 声母， 这些声母来源于某种上古汉语的齿音 (dental)声母。但是其中一些字的声母看起来是来源于上古汉语的舌根音(veral) 声母。本文指出了来源于上古舌音的切韵腭化音声母是属于切韵系统韵类中造成重纽区的字。根据声母清浊(k- 对 g-)和声调的条件，上古音中的舌根音声母的一些字发生腭化，而另一些则没有。我们找出了在切韵系统中象ts’je 对 gjie (重纽四等) 这样的腭化字。切韵系统中的腭化音声母跟有重纽四等的舌根音成互补状态。切音腭化音声母中来源于上古舌根音声母的字是可测的。
汉语普通话中元音对声母 r 的协同发音作用
Rongrong Liao 廖荣容
Coarticulatory effects that different vowels have on syllable initial consonant /r/ were examined. Spectrographic and LPC analyses were performed on all possible sequences of /VrV/ and /rV/, which yielded a total of 40 /VrV/ and 5 /rV/ sequences. 4 speakers were asked to produce the sequences. The results are as follows: (1) the F1 or F2 of /r/ covaries with that of the neighboring vowels positively, while the F3 is close to the F2 of /r/, and covaries with the F2 of the neighboring vowels through the F2 of /r/. (2) F2 of /r/ covaries more with F2 of the following vowels than with F2 of the preceding vowels. Either the anticipatory or carryover effect on F2 of /r/ extends beyond the boundary of /r/, but does not reach the central portion of the opposite vowel.
本文通过语音实验方法研究汉语普通话中不同元音对声母 r 的协同发音影响。汉语普通话中有四十种不同的“元音r元音”和五种不同的“r元音”组合。作者对由四位发音人所发的含这些语音组合的录音进行了语图分析， 并使用特定计算机程序 测量了这些语音的共振峰频率随时间变化的数值。对所的数据进行分析的结果如下：1. r 的第一共振峰和第二共振峰都因跟相邻元音的相应共振峰协同发音而产生频率变化。 r 的第三共振峰频率跟 r 的第二共振峰频率很接近，前者通过后者跟相邻元音的第二共振峰频率发生协同变化。 2. r 的第二共振峰因跟 r 后面元音第二共振峰协同发音而产生的频率变化比因跟 r 前面元音第二共振峰协同发音而产生的频率变化大。r 因跟相邻元音的协同发音，无论是先行的还是延迟的，反应在第二共振峰频率上的变化都一直延伸到 r 的边界以外， 但是达不到相邻元音的中心部位。
Hsiao-jung Yu 遇笑容
This study is an examination of the interrogatives used in Rulin waishi, a Chinese vernacular novel attributed to Wu Jingzi (1701-54) and one of the few novels in the body of traditional Chinese literature whose authorship has raised relatively little controversy among scholars and critics. Careful examination of the interrogatives used in the novel, however, reveals certain grammatical features that are common to the first thirty-two chapters of the novel but incongruent with corresponding features in the remaining twenty-three chapters, a phenomenon that suggests that the language of the earlier chapters preserves features of old Mandarin and Quanjiao dialect (Wu Jingzi’s own dialect) and that the language of the remaining chapters includes linguistic features of dialects other than Quanjiao. These “consistent inconsistencies,” especially in the distribution of the interrogative sentence patterns VP-bu-VP vs. ke-VP, argue strongly that Rulin waishi is very likely a work of dual (or multiple) authorship.
Evidence is presented for a rule, according to which word-final clusters of resonants and laryngeal stops develop into plain resonants in Sino-Tibetan, while the clusters of resonants and laryngeal fricatives are dropped. The paper contains more than a hundred comparisons between North Caucasian and Sino-Tibetan demonstrating this phonetic law.
这里为一条规则提供了证据。根据这一规则，在汉藏语中辅音与喉塞音词尾音丛(-r)演变成 -r, 而辅音与喉擦音词尾音丛(-hr) 则丢失。本文通过一百余条 North Caucasian 与汉藏语之间的比较证明了这一规则。
In this paper I examine a few syllable-sensitive phonological processes in Mandarin dialects, including partial reduplication, speech error, game languages, l-words and diminutive affixation. These processes target sub-syllabic constituents, therefore enable us to look into the internal structure of the syllable in Mandarin. I will show that first, the vowel and syllable-final glide or nasal form a constituent, i.e. rime; second, the constituency of the medial glide is not as determinate as commonly assumed; and third, to the extent that it is determinate, the two glides (i and u) exhibit asymmetric behavior with respect to their syllabic position.
本文详细检讨汉语中的一些和音节有关的音韵现象， 包括部分重叠，口误，反切语， 嵌-l词， 以及儿化等。 这些音韵现象涉及音节成分，通过分析，我们可以窥视官话音节的内部结构。本文的结论有三个。 第一，官话音节的元音和韵尾介音或鼻音组成一个成分， 即音基；第二，韵头介音的音节位置不易确定； 第三，即使韵头介音的位置能够确定， 介音 i 和 u 在音节中的位置不一定对称。
This paper defines ‘question’ as a category of discourse function, and suggests that it should not be confused with ‘interrogative’, which is a category of grammatical form. In this functional interpretation, a question is any information-seeking utterance, which does not have to take the form of an interrogative. The paper presents a critique of a traditional description of questions, and attempts a functional classification of questions in Cantonese. Five types of question are identified and discussed: Q: inform, Q: confirm, Q: restate, Q: clarify, and Q: display. The paper concludes by considering the advantages of this functional description of questions.
本文指出询问或提问 (question) 为一语言功能概念，而疑问句 (interrogative) 则为一句法概念。提问的目的是从听话人那里获取信息，这目的可透过不同的句型来达到，并不限于疑问句。文章讨论了用文法准则 (formal criteria) 来分析询问功能的不足之处，接着从语言功能的角度， 描述了广东话里的五种提问 (five types of question) 分别为(1)询问:提供信息 (Q: inform) , (2)询问:确实(Q: confirm), (3)询问 : 重复 (Q: restate), (4)询问：阐明(Q: clarify), (5)询问:展示 (Q: display). 文章最后解释了这功能分类法 (functional classification) 的优点。