Contemporary studies on the Min dialects
Edited by Pang-Hsin Ting 丁邦新 主编
Modern linguistic study on the dialects started with Luo Changpei’s work on the Xiamen dialect published in 1930. A detailed examination of the phonological relationship between the readings of individual characters in the Xiamen dialect and those in the Qieyun system, Luo’s book paved the way for subsequent research on Min dialectology. However, it was not until Tung Tung-ho’s study (1960) on the four South Min dialects that we began to have access to reliable descriptive data collected by modern field methods. His study provided fairly complete lexicons and included long stories, phonetically transcribed, instead of readings of individual characters.
Traditionally, the Min dialects are divided into two branches; namely, Southern Min and Northern Min. In 1963, Pan Maoding et al. classified the Min dialects into five subgroups: Eastern Min, Puxian, Southern Min, Central Min, and Northern Min. This subgrouping has gained further support from Li Rulong and Chen Zhangtai’s research findings (1991) and is now considered a standard classification.
With the publication of Norman’s “Tonal Development in Min” (1973), the field of Min dialectology embarked on a new era. For the first time, a proto-Min reconstruction system was introduced. Norman’s article has generated continuous debates on various issues, including the classification of the Shaowu dialect, and the additional tone in Kienyang. However, now that many of these controversies have been resolved, it is time to consider and propose a new reconstruction.
Ting (1979:265, 丁(1979:718) proposed that the Min dialects were probably derived from the main stream of the Chinese language during the Han dynasty, a claim he supported with further phonological evidence in 1983. In the same year, Norman (1979) argued that, based on lexical evidence, the earlierst stratum of the Min dialects could date back to the Han dynasty. This independent convergence in view lends great weight to the hypothesis.
Sung (1973) studied the difference between the literary and colloquial Min reading. Her article was the first attempt to systematically address the issue of linguistic strata in Min. Along the same line, Yang (1982) discovered that there were three strata in the dialect of Chaozhou and only two in Xiamen, Quanzhou, and Zhangzhou. It is important to note that while Norman was concerned with the pure lexical strata of double or triple readings of individual characters, Sung and Yang referred to entire layers of language based upon phonological analysis. In a more recent article, Chang (1996:59-66) argues that the Min dialects have four different layers formed in three time periods.
Ting (1988) suggested that the “Wu” dialect of the Southern Dynasties was the ancestor language of the modern Min dialects, and that the northern dialect of that time has eventually become the modern Wu dialects. He provided two kinds of evidence. The first type pertained to the colloquial substratum in the Wu dialects which showed a unique feature of the Min nature; the second type consisted of dialectal words which were used in the modern Min dialects but which were found in the Nanshi (南史) and the Wuge (吳歌). In connection, Zhou and You (1986:38) developed the idea of studying the origin of the Min dialects from the historical perspective of migration. Norman (1991:336-337) found out that Yang Xiong’s Fangyan and the commentaries by Guo Pu, a scholar in the Jin Dynasty, had preserved a large number of Wu (or Jiangdong) words, a solid piece of evidence which allowed him to assume that the language introduced to Fujian during the Han dynasty was some kind of a southern or Jiangdong dialect. These studies, again, came to the same conclusion, namely, that the modern Min dialects might have derived from the “Wu” dialecs of the Southern Dynasties.
Among the Chinese dialects, Min is a unique group characterized by various phonological and lexical features. As one of the major dialects spoken in Taiwan, the Min dialects have been the focus of research interests and activities. The products made by scholars in mainland China and the United States are euqally fruitful. It is now time to put together a collection of pertinent articles to give acknowledgment to the results of contemporary studies on the Min dialects.
In 1993, UC Berkeley received from the Chiang Ching-kuo Foundation of International Scholarly Exchange in Taiwan a generous grant to establish the Chao Yuen Ren Center for Chinese Linguistic. It is the purpose of the Center to conduct activities all in honor of this great scholar. Since 1994, as a way to encourage exchange of research experiences and findings, the Center has been organizing annual symposia focusing on different linguistic themes. Each year, a specific topic is identified and invitations for participation are sent to scholars in the United States, China, Hong Kong, Taiwan and Europe. The second annual symposium highlighted the studies of the Min dialects. The present volume of publication is based essentially upon the presentations delivered at the symposium held at UC Berkeley, in March 1995.
Since the Min dialects were most likely derived from Archaic Chinese of the Han dynasty, they naturally preserved old features no longer found in other dialects. Mei’s article focuses on the origin of several versatile particles in the Southern Min dialect spoken in Taiwn. He compares the dialectal material with both old colloquial data and other Min dialects. The use of some basic vocabularies in Modern Min is in fact a continuation from the Jiangdong dialect of the Southern Dynasties.
Yue’s article deals with the Min translation of Doctrina Chritiana, which was written probably in the last decade of the sixteenth century. She claims that the document consists of three levels of language: the colloquial, the literary vernacular, and the classical. In particular, the colloquial represents vocabulary items and syntactic structures unique to what we identify as the modern Min dialects. Ting (1992) differentiated dialectal history from the history of a dialect region. Yue’s study provides an apt example of how to use a small sample of material to illustrate the dialectal history of Min.
Chen (1993) published an article on the nature of entering-tone words in the Shaowu dialect. He argued that some irregular entering-tone characters had actually been derived from the diminutives. Historically, a final glottal stop was used as an affix to mark the diminutive. The segmental marking eventually rendered these diminutive words indistinguishable from the true entering tone words which always had a glottal stop. When the glottal stop disappeared from the language, the remaining components of both the entering tone words and the diminutives were represented by the same pitch contour. Norman took the surface value of the entering-tone contour of the diminutives in the Shaowu dialect and reconstructed *mh-, *nh-, *lh- in proto-Min, which has been proven to be incorrect. Chen’s current article is a continuation of his previous investigation. By comparing Min with its neighboring dialects. He argues that there are three kinds of diminutives: (1) the suffixal –jian, which is the original form and remains the most popular marker in the language, (2) the suffixal –er, which is introduced from northern Mandarin, and (3) the tonal alternation from which some of the Shaowu entering tone words are derived.
Robert Cheng established a data base entitled “Taiwanese-Mandarin lexicon” few years ago. By using the material in that data base, he is able to study two types of synonyms among pronouns. He discusses the historical origins of the synonyms, and analysis their grammatical, lexical and pragmatic particulars in early texts as well as in the modern dialects. Without the data base, one can hardly have a comlete picture of the pronouns, let alone a close understanding of their phonological, syntactic, semantic, and pragmatic behaivors.
Y. R. Chao (1968:13) claimed that “it is in matters of grammar that the greatest degree of uniformity is found among all the dialetcs of the Chinese language. Apart from some minor divergence, such as indirect object before direct object in the Wu dialects and Cantonese –for which Mandarin (like English) has the opposite order, and slight differences in the order of the negative in potential complements in some of the southern dialects, and so on, and apart from differences in suffixes and particels for which, however, fairly close equivalents can be set up between dialects, one can say that there is practically one universal Chinese grammar.” In recent studies, scholars have argued that, contrary to Chao’s observations, different dialects do show tremendous divergence in grammar. For instance, Zhu Dexi (1985) illustrated that the interrogative patterns were quite different among dialects. Yue-Hashimoto (1988) provided a comprehensive review of comparative dialect grammar during the last hundred years, and noted the direction for future studies. On the Min dialects, Yang (1991) published a descriptive grammar of the Southern Min dialects in Taiwan, giving us for the first time a complete picture of the language.
Lisa Cheng et al. study the morpheme hoo in the Southern Min dialect spoken in Taiwan, a marker that occurs in the causative, the passive, and the so-called double-object, dative and serial-verb constructions. Using the notion of “secondary predicate,” they offer a formal analysis of the argument structure of hoo that ties these constructions together. Following Chomsky’s approach, they assume that the Null Operator movement creates the secondary predicates. In Mandarin, the most equivalent word to the Min morpheme hoo is gei, which is not used in the causative at all. While it is not the intention of Lisa Cheng et al. to compare Min with Mandarin, their article nevertheless provides another solid example showing that dialectal grammars are not similar.
Lien’s article is a report on the sociolinguistic dimentions of comparative constructions in the southern Min dialect in Taiwan. He focuses on the correlation of the lexical and grammatical aspects of comparative constructions with social variables, including age, sex, education, and place of origin. For the age variable, he distinguishes six age groups, with a ten-year difference between two neighboring groups. For example, Group 1 is from 15 to 24, Group 2 is from 25 to 34, and so on. For the education variable, he divides the subjects into five classes who have attended: (1) primary school or are illiterate, (2) junior high school, (3) senior high school, (4) college, and (5) university. For the place of origin, he distinguishes subjects who are from the Hsinchu area from those who are not. In total he studies 150 subjects. The survey is evidently a large-scale project and most likely the first socialinguistic investigation on any grammatical issue ever conducted in Taiwan to date.
Tai has been working on his theory of iconicity and cognitive-based Chinese grammar for more than a decade. He (Tai 1992) pointed out that Chinese language offered a wealth of data for the study of classifiers and their implications for human categorization. Along the same line, his paper discusses a classifier, bue 53 “tail,” in Southern Min. He observes that the classifier did not exist in the Wei-Jin period when the classifier system had already been fully developed. It seems to have first appeared in the Tang Dynasty. The classifier has been preserved in the most conservative southern dialects only. Its cluster features for the categorization are: (1) animal, (2) long-shaped, (3) with a tail, (4) without legs, and (5) able to crawl or swim.
Chin-chuan Cheng (郑 1994. 1996) made significant contributions to quantitative studies of Chinese dialects and to the methodology of measuring the mutual intelligibility among dialects. Traditionally, dialect subgrouping has no formal means to express the inter-group distance or to illustrate closeness among related dialects. Cheng has designed a method to account for the dialectal correlation and the degrees of mutual intelligibility. For the first time, we have a precise way to express dialect distance on a numerical scale.
Finally, Ting’s paper deals with the reflexes of two Archaic Chinese initials in the Min dialects. He studies newly discovered Min materials, and offers two dinstinctive initials, namely, g and , instead of five as proposed by Norman, to account for proto-Min phonology. The article attempts to resolve one of the long standing problems in Min reconstruction and shows that proposing an alternative reconstruction of a new proto-Min system is possible.
As mentioned above, all the papers collected in this volume are revised versions of the papers presented at the School Annual Symposium of the Chao Yuen Ren Center for Chinese Linguistics. Thanks are due to the Chiang Ching-kuo Foundation of International Scholarly Exchange in Taiwan, which provided grants to establish the Center. I want to express my gratitude to Professor William S-Y. Wang, editor of the Journal of Chinese Linguistics, who has kindly consented to publish this volume as one of its monograph series. I am also indebted to Ms. Yifeng Wu, Assistant to the editor of the Journal. Without her timely and enthusiastic assistance, this volume would not appear in this shape now.
Tsu-Lin Mei 梅祖麟
This paper discuss the etymologies of the following grammatical particles in Southern Min: (1) plural suffix –n, (2) first person plural pronoun lan 3 ‘inclusive we’, (3) second person singular pronoun li 3, (4) interrogative lua 6, dzua 6 for the ‘how’ in ‘how far, how tall’, (5) auxiliary verb e 6 ‘can, able to’, (6) diminutive suffix a 3, (7) distal deictic particle hit 7 ‘that’, hia 1 ‘there’, and (8) locative particle ti 6. The discussion follows and refines previous work by Jerry Norman (1983) and the present author. For (1), it is shown that the Tsu-t’ang chi of 952 A.D. contains the expression 汝儂, plural ‘you’. From (4), it is proposed that the lua 6 in lua 6 kuan 2 ‘how tall’ is a fusion of 若 *lio/ 8 ‘how’ and 夥ua 6 ‘many’.
The vocabulary and the syntax of Doctrina Christiana en letra y lengua China of early seventeenth century are described in comparison with the same in two other texts written about the same time, the Jiajing edition of the Litchi Mirror Tale of 1566 and a Min Grammar Published in Spain in the early seventeenth century. The vocabulary discussed includes personal pronouns, interrogative forms, other pro-forms, demonstrative, negatives, function words (such as markers for comparison, for the benefactive/passive/causative, for the benefactive /comitative, and for the disposal form), modals, nominal suffix, as well as colloquial lexical items. The syntax part comprises affirmation and negation, the potential form, the resultative complement, measure expressions, the comparative, neutral question forms, as well as the attributive and the nominalizing construction. In terms of this study of vocabulary and syntax, three levels of style may be defined linguistically: a colloquial style identifiable as unique to the Min dialects, a literary vernacular (written) style approximating the standard spoken language of the time, and a classical style reflecting pre-Qin usage.
In this paper, I discussed the diminutives in Min dialects by comparing Min dialects with their neighboring dialects. I pointed out that there are three diminutives in Min. The first one is the jian suffix diminutive, which is the native form and is widely distributed in Min dialects. The second one is the er suffix diminutive, derived from the north, which can be divided into two different levels. One of these, with the er suffix, is mainly distributed in the Shaowu area of northwest Min. I call this suffix the new level, which shares the same phonetic features with the neighboring Gan dialects, such as the Lichuan and Nancheng dialects. In the Houlu dialect of Datian county, a subgroup of the southern Mindialects, there is another diminutive also derived from er suffix. It can be called the early level of the er suffix diminutive. This kind of er suffix diminutive was eroded in the most southern Min dialects, resulting in survival of the nasalized finals only in relic forms. The entering-tone characters of the Shaowu are relic forms of the third diminutive, formed by changing tones to express the meaning of the diminutive. These are two kinds of diminutives, the er suffix diminutive and the diminutive by changing tone, in the Shaowu dialect. These two diminutives also exist in some southern Wu, Hue and Gan dialects, the neighboring dialects of the Shaowu, which thereby shows the geographical continuity of this language feature.
Robert Cheng 郑良伟
The paper discuss two types of synonyms among pronouns: (1) those between three types of memory units, lexicon, morphological pattern, and regular syntactic structure, and (2) those between different language layers, pronouns with morphemes of old layer, new layer, and mixture of old and new. Quite a few pronouns as lexical entries are from the old layer in terms of colloquial reading or unclear etymology (e.g. lang5 ‘other’ lan2 ‘we inclusive’) or old day grammar (D-N chit-tah ‘here’; pat8-lang5 ‘other’). Function words used in regular syntax tend to be from the new layer *e.g. ki5-tha-e5 N, hou7-siong7 + VP). Function words in pattern tend to come from the old layer, whereas the content words can be from new or old. (sio-V, tak8-M).
The paper discusses the sources of synonyms, historical study of etymons in earlier Chinese character texts, and grammatical analysis of modern texts as well as the spoken language. Problems presented include: reference to the speaker and address by honorific and humble terms, the self-other pronouns, in addition to the first and second person pronouns; semantic classification of pronouns, of ‘self-other’ and ‘same-different’; different use of Chinese characters for Taiwanese morpheme, in earlier texts and modern ones.
All these concern the thorny problems of providing the syntactic, semantic, programmatic and phonological information to the Pronouns as entries in a Taiwanese Mandarin lexicon.
“人与己”或“自我和他人”在汉语是很常出现的观念对立。这个观念上的对立，很明显地反映不同记忆和认知单位的各层面里的语词上的对立：复合词里的词根（自杀：他杀），词汇单位虚词（“家己” （自己）：别人），词法里的虚词（sio照顾）句法里的回指规律（“家己无爱食，那会欲叫人食？” （自己不吃，怎么要叫人家吃？））。另外，语用中表达“被高举的中心人物” 和“被降低的局外人”人际关系的对立，这时“人” 可能指涉第一、第二、第三人称。
台语的“别人” 有时表达“人己” 的“人” ，有时又表达“异同” 的“异”。同时随着语意的精密化和表达明确化的需要，“别人” 的语意也可由几种新语言形式表达意义类似、却有细微差异的观念。这些演变反映在各种语词单位之间类义语的语意功能划分（人、别人、别个人、其它的人、另外彼个人）。其中量词的结构化导致表达明确化（仝人：无仝人，仝个人、别个人，仝彼个人：另外彼个人）。
"與", "與", "與": 台湾语言使役，与格及被动结构的句法
Lisa L.-S. Cheng 郑礼珊, C.-T. James Huang 黄正德; Y.-H. Audrey Li 李艳慧; C.-C. Jane Tang 汤志真
The morpheme hoo in Taiwanese has been commonly observed to occur in a number of constructions: the causative, the passive, and the so-called double-object, dative and serial-verb constructions. We offer a formal unifying analysis of the argument structure of hoo that ties these constructions together in a way that explains their clustering, to the exclusion of other constructions. We show that the double-object construction is a special instance of the canonical causative. The serial verb hoo is a causative embedded as a secondary predicate, and the so-called dative is a special case of such a secondary predicate. Finally, the passive is an ergativized version of the canonical causative, formed not by dethematizing the subject, but by turning its event complement into a secondary predicate. Following in part Feng (1995), Chiu (1995) and Chomsky (1980), we assume that the secondary predicates (in passives, datives, and the serial-verb structures) are syntactically created by Null Operator movement. We provide extensive arguments in support of this hypothesis and note some implications for the theory of argument structure.
台湾话动词 “与” hoo 的适用范围很广，包括致动、被动、双宾、与格和连动等句式。本文以一贯性的方式分析“与”字在这些句式里的参项结构，从而解释为何这些句式都能因使用“与” 字而“物以类聚” 。依作者之见，双宾句式其实是致动句式的一种特例，含“与”字的连谓结构则是一般的致动式包孕于母句之下充作次要谓语，而与格句式则又是连谓结构的一种特例。至于被动结构，我们认为它们来自致动句式的作格化。不过这种作格化和一般致动式的作格化不同。一般的作格化删除了致动动词的外向论元，但台湾话被动式则是将致动动词的内项化为次要谓语的结果。本文沿用feng (1995), Chiu (1995), Chomsky (1980) 等的分析，以“空运符移位” 来形成以上各类次要谓语，并提出相关证据来支持这种分析。最后我们指出这种分析对一般参项结构研究方面的理论意义。
Chinfa Lien 连金发
This paper is a preliminary report on sociolinguistic dimensions of comparative constructions in Taiwan Southern Min focusing on the correlation of lexical and grammatical aspects of comparative constructions with social variables. The linguistic types to be correlated include (1) the three tonal variants of the word 比, and (2) four comparative sentence patterns. We will examine in quantitative terms the correlation between the two types of linguistic behavior and social dimensions such as age, education, sex and place of origin. Issues of contact-induced change are also touched on.
A number of dialects are spoken in Fujian Province. Chen and Li (1983) and Li and Chen (1985) have chosen a few hundred morphemes, words, and phrases for comparison to establish a grouping for the 18 localities of Fuzhou, Gutian, Ningde, Zhouning, Fuding, Putian, Xianmen, Quanzhou, Yongchun, Zhangzhou, Longyan, Datian, Youxi, Yongan, Shaxian, Jian’ou, Jianyang, and Songxi. Once they have established the five subgroups of Eastern Min, Pu-Xian, Southern Min, Central Min, and Northern Min they have no formal means to express the inter-group distance except to reiterate the criteria for classification. This inability to express the degrees of closeness among related dialects is a weakness of the traditional language classification. In this paper two methods for quantification are discussed with respect to the dialects of the 18 localities. The first one deals with calculation of correlation. Utilizing the data given in Li and Chen (1985), Lu (1986) computed the correlation coefficients on the basis of sharing or lack of phonological and lexical-syntactic characteristics. A cluster analysis of the coefficients has yielded an affinity tree that is shown on a numerical scale ranging from 0.0 to 1.0, thus providing a way to express degrees of relatedness of these localities. The second method involves the calculation of dialect mutual intelligibility. The method proposed in Cheng (1992, 1994a, 1996) was used to derive the degrees of mutual intelligibility among these 18 localities. It first established sound correspondence patterns in the calculation process. The patterns that cover more morphemes can be thought of as general rules and should be assigned higher values than those covering fewer items. The syllable initials, medial, vowels, endings, and tones of all the items of the established patterns were given some values according to a weight scale established with this principle. The indices of Mutual intelligibility were then derived from these values. They provided a basis for establishing a grouping. The grouping looks much like what Li and Chen have established. But now we have a way to express dialect distance on a numerical scale.
福建省内方言多，陈章太、李如龙 （1983）以及李如龙、陈章太（1985）选取数百个单字、词语和词组来排比，建立福州、古田、宁德、周宁、福鼎、蒲田、厦门、泉州、永春、漳州、龙岩、大田、尤溪、永安、沙县、建瓯、建阳、松溪等18个地点的分区。闽东，蒲仙、闽南、闽中以及闽北五类方言片一旦建立起来，其间的关系和距离却没法表达，最多只能重复分类的基准来说明其分合。不能表达相关方言之间的关系程度是传统的语言分类法的缺点。 本文以这18个方言点来讨论两个计量方法。第一个方法是相关系数的计算，陆致极（1986）利用李如龙、陈章太（1985）的语料以音韵和词汇及语法的差异算出方言点之间的相关系数，接着把相关系数用聚类分析画出系属树形图，方言点之间的关系就在0.0到1.0的尺度上显示出来。第二个方法是方言沟通度的计算，用郑锦全（Cheng 1992， 1994a，1996）提出的方法算出这18个方言点闲的沟通度。计算程序是先建立语音的对当类型，类型所包括的字多的话就形成一般性的规则，因而给予比少数字的类型高的数值。每个类型的字的声母、介音、韵尾和声调都以根据这个原则建立的加权指数算出数值，沟通度就从这些数目演算出来。这18个方言点两两的沟通度就用来决定方言分区，所得的分区和李如龙、陈章太所提出的大致相同，但是我们能够在数值尺度上表达方言之间的距离。
古音声母*g和 *ɣ 闽语中的演变
Pang-Hsin Ting 丁邦新
According to the new material discovered in Shibo (石陂), a Northern Min dialect, this paper suggests that there were g and Ä initials in Archaic Chinese. The characters that belong to the xia initial (匣) in Archaic Chinese may be divided into two groups. One group, which originally combined with characters of the qun initial (羣), originated from Archaic Chinese g- and is preserved as g- in Shibei; the other group, which originally combined with characters of the yu initial (喻三), originated from Archaic Chinese Ä- and is preserved as ú- Shibei. Jerry Norman reconstructed five distinctive initials in his proto-Min system, namely, g, gh, -g, h, and Ä, which could be reduced to two, namely, g and Ä.