Mar 23, 2011 © Ulrich Theobald
There are different methods to arrange Chinese characters in a dictionary, namely the phonetical method, the graphic method, and the semantic method.
The phonetic method
Phonetic methods make it necessary to establish a phonetic transcription of characters. In modern dictionaries, modern transcription systems are used, in the People's Republic and in Western dictionaries mainly the Hanyu pinyin transcription
漢語拼音, in older Western dictionaries the Wade-Giles transcription
, and in older and in Taiwanese Chinese dictionaries the Zhuyin alphabet
注音字母 or the Gwoyeu Romatzyh transcription system
國語羅馬字. In traditional Chinese dictionaries there was a system of rhyme groups
with a fixed arrangement according to which characters had to be looked up.
The transcriptions using the Latin alphabet as well as Chinese symbols, like the Zhuyin
alphabet and its precursors of the qieyin method
切音, are explained in separate articles (see links). Information about the rhyme groups can be found in the articles to the particular dictionaries, see Qieyun
廣韻 and Jiyun
集韻. The last great dictionary using this system was the Peiwen yunfu
佩文韻府 from the Kangxi reign, but there is also a dictionary called Citong
辭通 from the Republican period using an arrangement of characters in rhyme groups. For the modern use, these dictionaries are to use only with difficult. Modern editions of the Peiwen yunfu
therefore include a graphic index. The reasons for the difficulty are that firstly the modern phonetic system is very different from that of Middle Chinese or early modern Chinese, and secondly that the user has to be familiar with the headlines of the rhyme groups, which consist of exemplary characters, like 東 for the level tone words with the endings [-uŋ] or [-ĭuŋ]. This rhyme group includes 34 groups of homophones that are headed by the words 東, 同, 中, 蟲, 終, 忡, 崇, 嵩, 戎, and so on. The homophones group of 蟲 [ɖʰĭuŋ] includes the characters or words 蟲, 沖, 种, 盅, 爞, 𦬕 and 翀.
The phonetic system was not only used for ancient dictionaries or indexes, like that of the biographic collection Guochao qixian leizheng
國朝耆獻類徵 (the index to which was created in XXX), but is occasionally used in dictionaries from the Republican period. A reminiscence of this method is to be found in the modern Reverse Chinese-English Dictionary
|Example of the phonetic index from the Guochao qixian leizheng 國朝耆獻類徵. The example shows personal names ending with the word nian 年. The word nian belongs to the rhyme group xian 先. The names of the persons are Heng Nian 恆年, Xu Yingnian 許應年, Henian 鶴年 (not a Chinese), Shen Danian 申大年, Wang Younian 王有年, etc. (from right to left). The small characters below the names indicate the chapter and the number of the scroll. Click to enlarge.
||Example from the dictionary Citong 辭通 from 1934. The upper picture is the beginning of the index, showing that it is arranged according to the traditional Guangyun rhymes (with the homophonic rhyme group 東, including the words 東, 同, 銅, 桐, etc. from right to left). The lower picure shows the first word in the dictionary (dong 東). It can be seen that the Citong is not only a rhyme dictionary containing very few words, but that it also arranged in reverse order, the keyword standing in the second position. Click to enlarge.
||Example from the Reverse Chinese-English Dictionary from. Click to enlarge.
|Example of a Wade-Giles index in the dictionary Da cidian 大辭典, Taibei: Sanmin shuju, 1985. Below the level of syllables, the characters are arranged according to traditional radicals (in the group chêng4, 亻口巾忄扌攵止灬爪...). Click to enlarge.
||Example of a Zhuyin index in the same dictionary, showing approximately the same phonetic groups (zhēng and zhèng). Click to enlarge.
||Example of a Pinyin index from the Hanyu da cidian 漢語大詞典. Here, too, the homophonic wors are arranged according to radicals. XXX unclear. Click to enlarge.
Characters with several different pronunciations (duoyinzi
多音字) pose problems if the dictionary is arranged phonetically. In modern dictionaries, such characters are listed twice, in both places of alphabetical order, and an according instruction is given in the alphabetical index as well as directly under the characters. Older Western dictionaries list the character under the more frequently used pronunciation, which is in many cases not the contemporary standard. This is also often the case if the dictionary is arranged graphically: The common pronunciation is presented firstly, and only then less common or frequent sounding(s).
A graphical arrangement of characters is traditionally made by the use of radicals
. Virtually all character dictionaries, even those in which the characters are arranged phonetically, have one or several graphical indices because the pronuncation of a rare character can not be known, or someone might have forgotten the pronunciation or might not know the transliteration system used (like the Zhuyin
alphabet). Many Chinese, especially those from the south, do not use the Hanyu pinyin
system correctly and are therefore in need of other tools for looking-up a character or a word.
Dictionaries with characters arranged according to the Hanyu pinyin
system have to consider that the letter ü
is put after the letter u
(as a graphic variant), and the syllable [nuan] nuan
therefore follows the syllable [ny] nü
, inspite of the fact that the sound of [y] is very different from that of [u]. Within one syllable, the words are still arranged phonetically according to the tone pitches. Below the level of tone pitches, characters being graphically "relatives" (using the same phontic element) are often standing side by side, like
- 麗儷邐酈驪鸝. (Mathews' Chinese-English Dictionary, including words of all tone pitches except those with the entering tone).
Inside the groups of phonetic relatives (in the example below: 麗), the characters are arranged according the the stroke number of radicals, in raising order:
- 麗 (0),
- 儷 (2),
- 邐 (7, radical 辵),
- 酈 (7, radical 邑),
- 驪 (10),
- 鸝 (11).
Modern dictionaries prefer the number of "brush" strokes as criterion for arrangement, like
- 礼 (5),
- 李, 里 (7),
- 俚 (9),
- 逦, 哩,浬, 悝, 娌 (10),
- 理 (11),
- 锂 (12),
- 鲤 (15),
- 澧 (16),
- 醴 (20),
- 鳢, 蠡 (21). (Xiandai hanyu cidian 現代漢語詞典, 5th ed.)
If not only characters are to be arranged, but words, there is the complex question of either strict alphabetical arrangement or arranging the words according to the first syllable or even the character of the first syllable. In the first case words beginning with the same character can be dispersed through the register so that it seems that a word looked for is not included in the index. On the other side, a strict alphabetic order regardless which characters are used or where the end of the syllable is seems to be less complex.
The graphic method
- jiālóng 家隆
- jiālòng 夾衖
- jiǎlóng 假龍
- jiàlóng 駕龍
- jiǎlóu 假樓
- jiǎlóuluó 迦樓羅
- jiālù 夾路
- jiālù 家鹿
- jiālù 家祿
- jiālù 嘉露
- jiāshēng 家生
- jiāshēng 家聲
- jiāshēng 嘉生
- jiāshēng 嘉牲
- jiāshēng 嘉聲
- jiāshēng 挾生
- jiāshèng 佳勝
- jiāshèng 家乘
- jiāshèng 嘉勝 (Victor H. Mair (2003). An Alphabetical Index to the Hanyu Da Cidian. Honolulu: University of Hawai'i Press.)
The radical system as such was invented at the end of the Han period
漢 (206 BCE-220 CE), but Chinese scholars were long before aware that there was a limited amount of classifiers according to which all characters could be grouped in semantic fields. Not all radicals are quite useful, and in many characters they play only a classifying part in a graphical sense, without referring to the semantic of the characters. Xu Shen 許慎, compiler of the dictionary Shuowen jiezi
說文解字, has established a system of 540 radicals. This system has, with some minor changes, survived until the Ming period
明 (1368-1644) when it was replaced by a system of 214 radicals, which was perpetuated in the character dictionary Kangxi zidian
Most radicals indicate a semantic field:
- 心 "heart, feeling" for
- 必 "necessary",
- 忍 "to endure",
- 志 "will",
- 忘 "to forget",
- 忙 "in haste",
- 念 "to think of",
- 怕 "to fear",
- 性 "character",
- 情 "feelings, love",
- 恥 "to feel ashamed",
- 愛 "to love"
- 手 "hand, action" for
- 打 "to beat",
- 批 "to judge, to criticize",
- 折 "to cut",
- 承 "to receive",
- 拜 "to beg",
- 拳 "to hold high",
- 掌 "to control",
- 摯 "to hold"
- 金 "metal" for
- 釜 "kettle",
- 銀 "silver, money",
- 銅 "copper",
- 釘 "nail",
- 銳 "sharp",
- 鋒 "blade, tip of a knife",
- 鑒 "mirror"
Others are often purely graphical, like 一 "one", 乙 (a bent stroke), 二 "two", or 亠 "head, above". Most radicals are still used as a word by themselves, but others are obsolete since long and have the only function of radical. Radicals not used as words (if they ever had been word at all!) are 冂 "wilderness", 匚 "box", 卩 "seal", 厂 "cliff", 夊 "to walk", 宀 "roof", 尢 "curved leg", 广 "house", 彐 "pig head", 爿 "bed", 疒 "sick", and many more. A few of them have been reintroduced as simplified characters
, but with totally different meanings, like 厂 as abbrevation for 厰 "workshop" or 广 as abbreviation for 廣 "broad, wide".
Some radicals change their appearing in the regular script depending on their position in the character, like:
- 心 "heart" (below) or 忄 (left)
- 手 "hand" (below) or 扌 (left)
- 水 "water" (below) or 氵 (left)
- 火 "fire" (left) or 灬 (below)
- 犬 "dog" (below) or 犭 (left)
- 衣 "clothing" (below) or 衤 (left)
|New radicals in the Cihai 辭海, 1977 ed.|
Modern dictionaries in Taiwan still use this system, yet with the simplification of characters in the People's Republic, a change in the radical system was undergone (see picture). There is, nonetheless, still no authoritative radical system for the simplified characters, and each dictionary has its own approach. Radicals are since more used as graphical elements and less as semantic indicators. Some systems make it possible to look up for a character under different possible radicals, like the latest editions of the Xiandai hanyu cidian
現代漢語詞典. Most modern dictionaries from the People's Republic also discern between the different shapes of a radical, so that the words related to "water" have to be looked after either under the radical 水 or under the radical 氵.
In dictionaries using a radical system (however it looks like) the characters, being a word or the first syllable of a word, are arranged according the number of the brush strokes left over AFTER subtracting the number of the brush strokes of the radical. This is important because a radical can have a different amount of strokes depending on its position inside the character (水 at the top has four strokes, at the left, as 氵, only three). In most dictionaries the characters with the same number of strokes are arranged according to the five shapes of the brush strokes of the remaining part of the character.
Such graphical arrangements are also used in indices of all types of dictionaries and handbooks, even if the index as such is arranged phonetically and not according to radicals. Virtually all Chinese indices use the brush-stroke method to sort characters, even if it is not explicitly said so. This method sorts characters according to the shape of the first and the second brush stroke of the characters, in many indices only according to the first. There are five different brush stroke shapes (bixing
筆形) as a criterion:
- horizontal 一,
- vertical 丨,
- bend to the left 丿,
- dot 丶,
- and hook 乙.
Indices using the first two strokes arrange the characters sort the heading in the sequence一一, 一丨, 一丿, 一丶, 一乙, 丨一, 丨丨, 丨丿, 丨丶, 丨乙, and so on.
Some dictionaries make themselves totally free from traditional characters and provide a method that is purely graphical, like the Zhongwen zipu
中文字譜. Lin Yutang and Liu Daren have developed their own graphical systems which are only used in their dictionaries.
The largest common modern dictionaries arranged according to the radical system are the Hanyu da cidian
漢語大辭典, a word dictionary, and the Hanyu da zidian
漢語大字典, a character dictionary.
A very practical method which is not easy to learn is the Four-Corners-Indexing method
四角號碼. This method gives each character a four-digit number according the the shape of the “brush” strokes on the four (virtual) corners of the character. To become proficient in this method, some training is necessary, yet the presence of four-corner-indices in many dictionaries and handbooks shows that it can be used as a quick method to have easily access to characters and words.
Similar systems have been created by Lin Yutang and Liu Daren, and the editors of the Harvard-Yenching index series. The Cangjie system of character input for text processing is also working with a method of graphical analysis free from the traditional radicals and semantic elements.
The semantic method
The semantic arrangement of characters goes back to the tradition of gloss book dictionaries that assemble words of similar meanings. From simple lists like in the Han period gloss books Erya
爾雅 and Shiming
釋名 and the Guangya
廣雅 from the Wei period, these kinds of books developed two branches: Firstly, dictionaries specialized on disyllabic words, like the Pianya
比雅 and Dieya
疊雅, and secondly, a branch that grew into the direction of encyclopaedias (leishu
類書). To look up for a word in an encyclopaedia is an even more complicated matter than in a rhyme dictionary. Modern editions of encyclopaedias therefore have all indices. The last traditional encyclopaedia is the Gujin tushu jicheng
古今圖書集成 from the Kangxi reign. The entries in modern encyclopaedias like the thematic volumes of the Zhongguo da baike quanshu
中國大百科全書 are arranged phonetically, but there are also indices in each volume, as well as a general index volume.
One example from the Confucian Classic Erya
shall demonstrate how the gloss books looked like:
"Being a statute", "regulating", "levelling", "being standard", "punishing", "modelling", "serving as carpenter's square", "being common", "regular", "musically tempered", "straight like a lance", "employed", "ordered", all this means "what is the rule".
- 柯、憲、刑 、範、辟、律、矩、則，法也。
"Straight like a trunk", "being a law", "punishing", "modelling", "governing", "musically tempered", "serving as carpenter's square", "being standard", all this means "making a common level".
Cao Xianzhuo 曹先擢 (1988). "Gujinzi 古今字", in Zhongguo da baike quanshu 中國大百科全書, Yuyan wenzi 語言•文字 (Beijing/Shanghai: Zhongguo da baike quanshu chubanshe), 97.
Chen Fu 陳紱 (1993). "Benzi 本字", "Chongwen 重文", "Gujinzi 古今字", "Huiyi 會意", "Jiajie 假借", "Jiezi 借字", "Leizengzi 累增字", "Tongjia 通假", "Yitizi 異體字", "Zhuanzhu 轉注", in Zhongguo xiaoxue jiaoxue baike quanshu zong bianji weiyuanhui Yuwen juan bianji weiyuanhui 中國小學教學百科全書總編輯委員會語文卷編輯委員會 (ed.), Zhongguo xiaoxue jiaoxue baike quanshu 中國小學教學百科全書, Yuwen 語文 (Shenyang: Shenyang chubanshe), 151-154.
Wang Jihuai 王吉懷 (2000). "Zai lun Dawenkou de taoke 再論大汶口的陶刻, Dongnan wenhua 東南文化 2000/7, pp. 6-14.
Xu Zhongshu 徐中舒, ed. (1990). Jiaguwen zidian 甲骨文字典 (Chengdu: Sichuan cishu chubanshe).
Zhang Zhenglang 張政烺 (1988). "Bafen 八分", "Caoshu 草書", "Guwen 古文", "Lishu 隸書", "Niaochongshu 鳥蟲書", "Xingshu 行書", "Zhenshu 真書", "Zhouwen 籒文", "Zhuanshu 篆書", in Zhongguo da baike quanshu 中國大百科全書, Yuyan wenzi 語言•文字 (Beijing/Shanghai: Zhongguo da baike quanshu chubanshe), 10, 31-32, 257-258, 303-304, 430, 515-516, 538, 542-543.
Zhou Zumo 周祖謨 (1988). "Hanzi 漢字", in Zhongguo da baike quanshu 中國大百科全書, Yuyan wenzi 語言•文字 (Beijing/Shanghai: Zhongguo da baike quanshu chubanshe), 195-199.
Zhu Xiaojian 朱小健 (1993). "Bihua 筆畫"， "Bishun 筆順", in Zhongguo xiaoxue jiaoxue baike quanshu zong bianji weiyuanhui Yuwen juan bianji weiyuanhui 中國小學教學百科全書總編輯委員會語文卷編輯委員會 (ed.), Zhongguo xiaoxue jiaoxue baike quanshu 中國小學教學百科全書, Yuwen 語文 (Shenyang: Shenyang chubanshe), 156.
Zhongguo shehui kexue yuan kaoguxue yanjiusuo 中國社會科學院考古學研究所, ed. (1984). Yin-Zhou jinwen jicheng 殷周金文集成 (Shanghai: Zhonghua shuju).