Origin and Types of Chinese Characters

The invention of Chinese characters is traditionally attributed to Cang Jie 倉頡, a minister of the mythological Yellow Emperor 黃帝. He created Chinese characters as an imitation of the footprints of birds. The book Daodejing 道德經 that is attributed to the philosopher Laozi 老子 says that in oldest times people used knotted cords (jiesheng 結繩) as a system of remembering (quite similar to the quipu cords used by the Incas). The Eight Trigrams (bagua 八卦, see Yijing 易經) invented by the mythological ruler Fu Xi 伏羲 are also sometimes seen as a kind of primitive script. The same is true for the River Chart (Hetu 河圖) and the Inscription of the Luo (Luoshu 洛書).

Unfortunately there is no stage known between the age when characters were used to signify clan names or personal names, and the time when the script appeared in a full stage on the oracle bones (jiaguwen 甲骨文) and bronze vessels (jinwen 金文) of the Shang period 商 (17th-11th cent. BCE). The texts of the oracle bone inscriptions are highly technical and therefore represent only a narrow lexicon. About 2,000 characters are known from the oracle inscriptions, but the meaning of a lot them is indefinable. Part of the problem is that characters at that time were not yet standardized.

The script was still in a state of experiment, althought it is very clear that new characters were systematically developed. Some clan insignia or primitive symbols are also to be found as carvings in pottery, or as stone inscriptions (shike 石刻).

The creation of new characters

Chinese characters have a long history and therefore also change over time. Some characters were newly invented, while others became obsolete. Some changed their meaning, while others changed their appearance. Outdated characters were called guzi 古字 "ancient characters", while new creations were given the name jinzi 今字 "modern characters" or houqizi 後起字 "later creations". Ancient characters were often used to write several different words, because the thesaurus of characters was still quite modest. Later on new characters were created for the use of different words or equally pronounced words with a different meaning.

莫 /mˤak, MC mò/, for instance, was the proper character for the word "dawn", but it was also used as a negative pronoun "nobody". Later on, a new character was created for the word "dawn", namely 暮 (MC mù), while the original form 莫 was exclusively used as a negation. 賈 /Cqˤaʔs, MC jiǎ/ was the proper character for "to trade, to marketize", but was also used for the word "price". Later on a new character was created for the word "price", namely 價 (MC jià). 衰 /sruj, MC shuāi/ was the proper character for "weak, soft, exhausted", but was also used for the words "mourning clothes" and "rain coat". Later on two new characters were created, namely 縗 (MC shuāi) or 𧟄 (MC ráng) for "mourning clothes" and 簑 (MC suō) or 蓑 for "rain coat", to write the two other meanings.

The process of change from ancient to new characters had three different ways.

1) The ancient character was used for a different, yet equally pronounced, word, while a new character was created for the proper word:

2) A new character was created for the different meaning with close or equal pronunciation, while the proper character continued being used for the actual meaning:

3) The old character was used as a loan character for many different, equally pronounced words, for which later new characters were created to make comprehension easier.

There are whole character families (zizu 字族) that are a mixture of a signific-phonetic characters (xingsheng zi) and signific-signific characters (huiyi), like

Because there was no standarized script in ancient times, totally different characters might be used for one and the same word:

In some cases two different steps of creation can be observed, like:

Apart from these examples, the Chinese lexicon includes a lot of specialized words that are only used in a very narrow field, like geographic names as

botanical and zoological terms, as

and many other special terms with characters like

沐	"washing the hair"
沬	"washing the face"
浴	"washing the body"
澡	"washing the hands"
洗	"washing the feet"
臲鼿	"jumpy, jittery, worried"
衚衕	(simple 胡同) "hutong, dwelling quarter in Beijing"
鞦韆	(simple 秋千) "swing"
	a talismanic variant for 岩 "cliff" (and the character with the highest number of brush strokes, namely 51)

The six types of characters liushu 六書

The term liushu 六書 "six types of characters" is first mentioned in the Confucian Classic Zhouli 周禮, as a general word for "the six (arts of) reading and writing" (see liuyi 六藝). Liu Xin 劉歆 (d. 23 CE), in his treatise Qilüe 七略, enumerates the six types of characters as xiangxing 象形 "illustrating a shape", xiangshi 象事 "illustrating an affair", xiangyi 象意 "illustrating an idea", xiangsheng 象聲 "illustrating a sound", zhuanzhu 轉注 "mutual comment" and jiajie 假借 "wrongly borrowing".

The first four are actually not necessarily types of characters but rather elements in characters. The last two are types of character use, the zhuanzhu type as exchangeable characters with similar or equal meaning, the jiajie type loan characters for words with identical pronunciation. In the respective ancient literature, no examples are provided for these six types, so that the exact meaning of them is not clear.

Zheng Zhong 鄭眾 (d. 83 CE), author of a commentary on the Zhouli 周禮 (Zhouguan jiegu 周官解詁), changed this listing to xiangxing 象形 "illustrating a shape", huiyi 會意 "combined meaning", chushi 處事 "dealing with an affair", xiesheng 諧聲 "harmonized sounds", zhuanzhu 轉注 "mutual comment" and jiajie 假借 "wrongly borrowing". This new list was not a new concept of Liu Xin's older list, but only a renaming.

The most famous concept of the six types of characters has been established by the Later Han-period 後漢 (25-220 CE) scholar Xu Shen 許慎 (c. 58- c. 147) in the preface to his dictionary Shuowen jiezi 說文解字. He uses the following pattern of character types: xiangxing 象形 "illustration of a shape", zhishi 指事 "pointing at a matter", huiyi 會意 "combined meaning", xingsheng 形聲 "shape and sound", zhuanzhu 轉注 "mutually commenting", and jiajie 假借 "wrong borrowing".

Xiangxing 象形

Xiangxing 象形 "illustration of a shape" is an ideograph of objects, either simple (chun xiangxing 純象形) or complex (heti xiangxing 合體象形 "compound", or bianti xiangxing 變體象形 "changed"), like objects in nature (the left character is the seal script shape, the right its modern standard script counterpart)

日 "sun"	月 "moon"
山 "mountain"	水 "water"
木 "tree"	气 "vapour, breath"
雨 "rain"	云 "cloud"

different animals and plants like

馬 "horse"	羊 "sheep"
鳥 "bird"	烏 "crow" (a bird whose eyes cannot be seen)
竹 "bamboo"	米 "grain"

or parts of the body like

手 "hand"	首 or 頁 "head"
目 "eye"	自 "nose" (soon loan word for "oneself")
耳 "ear"	口 "mouth"
面 "face"	毛 "hair (not on the head)"
肉 "flesh, meat"	牙 "teeth"

or various objects and tools of man's environment like

田 "field"	耒 "egg (agricultural implement)"
糸 "thread, fathom"	韋, 革 and 皮 (different types of skin and leather)
車 "cart"	舟 "boat"
弓 "bow"	矢 "arrow"
豆, 鼎, 豊, 豐, 斝 and 鬲 (different types of vessels)	缶, 壺, 酉, 皿 (different types of jugs and jars)
門 "gate"	戶 "house"
衣 "cloth, robe"	网 "net"
玉 "jade" (three joint pieces)	文 "pattern, script"

This group also includes symbols of figurative meaning, like

交 "exchange" (a picture of crossed legs)	出 "going out" (originally a foot coming out of a compound)
步 "to go" (two steps)	立 "to stand, to erect" (a man above a horizontal line, as the earth)
至 "to arrive" (an arrow hitting the target)	折 "to cut" (a hand holding an axe; is probably of the huiyi type; compound)
天 "Heaven" (large 大 man 人 with indicated head)	示 "heavenly bless" (coming down from Heaven)
見 "to see" (rays emanating from the eye 目)	言 "to speak" (waves coming out of the mouth 口)
王 "king" (the one 丨 combining the trinity 三 of Heaven, Earth and man; actually the pictogram of an axe as the symbol of power )	"to obtain" (a hand and a shell, expressing value; the component 彳 "proceed" of 得 was added later; compound)
止 "to stop" (image of a foot")	行 "to walk" (image of crossroads)
高 "high" (image of a palace tower)	長 "long" (image of hairs)
力 "force" (image of a plow)	豐 "abundant" (image of a vessel containing pieces of jade)

Zhishi 指事

Zhishi 指事 "pointing at a matter" is the visualization of an idea, either simple (chun zhishi 純指事) or complex (heti zhishi 合體指事 "compound", or bianti zhishi 變體指事 "changed"), like

上 "above" (the plain 一)	下 "below"
本 "root" (of a tree 木)	末 "branch"
刃 "blade" (of a knife 刀)	甘 "sweet" (something 一 in the mouth 口)
帀 "circumference" (a turned ㄓ, i.e. 之 "to proceed")	乏 "deficient" (opposite of the turned character 正 "correct")
𠤎 "change", a turned 人 "man"	司 "court official, minister, administrator" (a turned 后 "lord", later "queen, empress")
母 "mother" (wife 女 with large breasts)	ㄊ "to discard" (a child 子 head down)

Characters of this category are either basic characters altered by an additional stroke, or charactes turned around or mirrored.

Huiyi 會意

Huiyi 會意 "combined meaning" is a combination of two or more characters to a new one, like

武 "war" from 戈 "halberd" and 止 "foot"	信 "trust" from 人 "man" and 言 "speech" (Xu Shen's analysis of 信 [ɕin] xin is wrong. Later scholars have found out that it is actually a combination of 言 "to say" and the phonetic 千 [tɕʰiɛn] qian.)
喪 "funeral" from 哭 "weeping" and 亡 "gone, dead"	旦 "dawn" from 日 "sun" and 一 the horizon
公 "public" from 八 "to separate" and ㄙ "private"	折 "to cut" from a form of hand 手 and an axe 斤 (斧)
囚 "prisoner", from 人 "a person" in 囗 "a fencing"	矦(侯) "lord" (a man shooting an arrow 矢 against a wall 厂)
男 "man" ("field" 田 and "plough" 力 [today "force"])	尾 "tail" ("body" 尸 and "hair" 毛)
畺 "borderline" (fields 田 with demarcation lines 一)	黍 "lacquer" ("plant" 禾 [today: grain plant] from which a liquid 水 emerges)
安 "peace" (a woman 女 under the roof 宀)	家 "family" (a pig 豕 under a roof 宀)
好 "good; to love" (girl 女 and boy 子)	里 "village" (fields 田 and earth 土)
爨 "to cook" (two hands 𦥑 holding a pot, below two hands 廾 that add firewood 木 to the fire 火)	黃 "yellow" (fields 田 are emanating 光 their colour)
赤 "red" (large 大 and fire 火)	黑 "black" (flames 炎 and chimney 囱)
鬥 "to fight" (two men with a weapon in the hand face each other)	食 "to eat" (to assemble 亼 fragrant food 皀)
取 "to take" (hand 又 and ear 耳)	葬 "burial" (corpse 死 in grass 茻)

Some huiyi-type characters also include a phonetic element, like

化 huà	"to change" ("person" 人 and turned person huà 𠤎)
政 zhèng	"to govern" (zhèng 正 "rectify" with a "whip" 攴,攵)

Many huiyi-type characters are easy to remember and are therefore very popular, not only in practice, but also to demonstrate how Chinese characters are produced.

Quite some huiyi characters are muliplications of characters or of their variant, like:

Huiyi characters have basically no phonetic element, but there are also some examples of xingsheng "shape-sound" characters in which the phonetic part is also used as a bearer of a meaning (see next type).

Xingsheng 形聲

Xingsheng 形聲 "shape and sound" is a combination of a signifying part (radical, bushou 部首) and a phonetic part, like

This is an advanced type of character that perhaps came up relatively late in the history of the invention of characters. It is, nevertheless, that type of character that is used most: 80 or even 90 per cent of characters belong to the xingsheng type. Some characters newly created according to this pattern were even to replace older, more pictographic types, like

In this list of oracle bone characters it can be seen that the fifth character already bears the phonetic 凡 (for its shape, see the upper part of character No. 8), before it became a regular part of the character (from No. 10). The characters are arranged chronologically. No. 17 is the standard seal script character for "phoenix". In this shape, the phonetic part has become very large and even includes the signific part "bird". From Xu 1990: 394-395, 427-428.

In this list of oracle bone characters it can be seen that the second character already bears the phonetic 奚. No. 10 is the standard seal script character for "hen, cock". The original form of the cock (No. 1) is step by step contracted to that of a "sparrow".

The change of phonetics in the last 2,000 years has reduced the feeling of users that characters have indeed a phonetic aspect. The modern reading of characters does often not have the slightest homophony with the phonetic parts once used to express the sounding. Even in Archaic Chinese the phonetics have not exactly the same pronunciation, but only a quite similar one. The creators of new characters did furthermore not necessarily use homophones (word of equal sounding) for the phonetic part of the new character, but often only "homoiophones" (words sounding similar). The phonetic part of characters was therefore already during the Han period only an approximate indication of the pronunciations, and by no means a scientific transcription.

Quite a few xingsheng type characters are consisting of phonetic elements that also contribute a meaning to the character. These characters are thus a combination of the huiyi and xingsheng types, often doubling the signific element.

In characters of the xingsheng type, both signific (shengxing 省形) and phonetic (shengsheng 省聲) might be abbreviated, like:

There are also different possibilities to render a certain sound with the help of different phonetics, or the meaning by different significs. More information can be found below, in the paragraph about character variants.

Zhuanzhu 轉注

Zhuanzhu 轉注 "mutually commenting" are characters of similar meaning and similar pronunciation, like

Some scholars explain this type as a kind of tautology like

An even narrower definition is that both characters are to have the same radical or signific part as a basic condition for their mutual exchange. Yet there are also examples of zhuanzhu characters with the same phonetic and different signific parts, like

Jiajie 假借

Jiajie 假借 "wrong borrowing" are characters that are borrowed for a word of the same pronunciation and for which no character exists. Xu Shen gives the examples

Yet modern scholars say that these two examples are only extensions of the original meaning, and not loan characters. Better examples are

Loan characters (jiezi 借字) also occur if the writer forgot or did not know the proper character. Some original characters (benzi 本字) have later often become obsolete (like 垔 as the original character of 堙 or 陻). In ancient writings a character often stood for very different things, like 齊 "equal, all" for 齋 "pious, chaste; to fast, to purify", 劑 "medicine, dose", 臍 "navel", 躋 "to rise, to ascend", 韲 "powdered, seasoning" and 薺 "water chestnut; caltrop; shepher's burse". The last few characters have only been created during the Han period or even later - they are new inventions.

Similar examples can be shown in almost all grammatical particles or pronouns, like

If a character was borrowed for another meaning, it was common that a new character was created for the original meaning of the word, like

The addition of a phonetic (radical) to create a new character out of a loan character, is called "increased character" (leizengzi 累增字). In such characters, the new addition of a radical is superfluous because the basic meaning of the character was indicated by a signific part, like:

Jiajie characters were already used in an early stage, which shows that the Chinese script had clearly a phonetic character and was not only used as a symbol script (ideographic). The oldest examples of loan characters appear in bronze inscriptions of Archaic Chinese, where the conjunction and auxiliary verb qiě 且 of Classical Chinese stands for zǔ 祖 "forebear", cái 才 (in Classical Chinese meaning "then, exactly, material, talent") for zài 在 "to reside in", or zhà 乍 (nearly obsolete in Classical Chinese) for zuò 作 "to create, to produce".

A verse in the Shijing 詩經 "Book of Songs" goes Zhi si shi mi ta 之死矢靡它, where shǐ 矢 "arrow" is used for the character shì 誓 "to pledge". In later times, loan characters were more subtile, like in Xunzi's 荀子 sentence mu zhi zhong sheng, ruan yi wei lun 木直中繩，輮以為輪, where róu 輮 "exterior rim of wheel" is used for róu 煣 "to bend wood over fire". Loan characters can have negative results for the understanding of a text, like in the Lunyu 論語 "Confucian Analects", where it is said gui Kongzi tun 歸孔子豚 "[he] sent a present of a pig to Confucius" (and not "sent Confucius to the pigsty" or else). Gui 歸 (literally "to return; to send sth. to where it belongs") is in this case used for kui 饋 "to feed, to present food".

Yet such characters and words are also often used to play with words, because in most cases the pronunciation of both characters is similar or even identical. Loan characters were very popular during the Warring States and Han periods, when the Chinese script was still in a state of development and standardization. Later on, when the script had become more regular concerning the use and shape of characters, the replacement of one character by another was often misleading. Commentaries to ancient books therefore often deal with the meaning, pronuncation and use of particular characters.

Some books regularly replace characters (a method called tongjia 通假 oder tongjie 通借 "general borrowing") with loan characters, like zao 早 "early" rendered as sao 蚤 (actually "flea"), xin 信 "to trust" for shen 伸 "to reach out", or guan 貫 "string", for guan/huan 宦/官 "official, minister".

Even today, there are quite a few loan characters, like lüyou 旅游 (游 with 氵=水 "water", actually "to swim") instead of 旅遊 (遊 "to travel", with 辶=辵 "to walk"). For the process of simplifying characters in order to increase literacy, a lot of loan characters are used, for example, gǔ 谷 "valley" for gǔ 穀 "grain", or jǐ 几 "table" for jǐ 幾 "some, any".

A very special kind of loan character, and in some cases loan words, are taboo names (bihuizi 避諱字) that are used in personal names in order to avoid a certain name, mainly part of an emperor's name, or the name of one's ancestor. The origin of taboo names is probably during the Qin period 秦 (221-206 BCE), when the name for the first month was changed from zhengyue 正月 to duanyue 端月 because Ying Zheng 嬴政 was the personal name of the king of Qin and First Emperor 秦始皇帝 (r. 246-210 BCE). The personal name of the founder of the Han dynasty was Liu Bang 劉邦 (Emperor Gaozu 漢高祖, r. 206-195 BCE). In order to avoid his personal name, the word bang "land, territory" 邦 was systematically changed to guo 國, actually the designation for a walled city. Emperor Jing 漢景帝 (r. 157-141 BCE of the Han dynasty had the personal name Liu Qi 劉啟, for which reasons the word qi 啟 "beginning" was regularly replaced by kai 開 "to open". Similarly, chang 常 "regular, standard" was chosen for heng 恆 "permanent, eternal", because it was the personal name of Emperor Wen 漢文帝 (r. 180-157 BCE). This regulation had even an impact on the beginning of the book Daodejing 道德經, which is since quoted as dao ke dao, fei chang dao 道可道，非常道, instead of dao ke dao, fei heng dao 道可道，非恆道, as attested in earlier sources. Emperor Wu 梁武帝 (r. 502-549) of the Liang dynasty 梁 (502-557) was called Xiao Yan 蕭衍, therefore the scholar Wang Yan 王衍 was called by his courtesy name Wang Yifu 王夷甫, in order to avoid the dynastic name. Li Shimin 李世民, the personal name of Emperor Taizu 唐太宗 (r. 626-649) of the Tang dynasty 唐 (618-907) made it necessary to write the character 世 without the right stroke , or 代 "age" instead of 世 "generation", and 人 "man" instead of 民 "people" (or without the last stroke: ). The name of Li Yuan 李淵, father of the dynastic founder, required a change of the character 淵 "depth" to 泉 "source", or 深 "deep". Similarly, the name of Zhao Jing 趙敬, father of Zhao Kuangyin 趙匡胤 (Emperor Taizu 宋太祖, r. 960-975), the founder of the Song dynasty 宋 (960-1279), made it nessecary to replace 敬 "to rever" by 恭 "to venerate" or 嚴 "to take serious, and to replace jing 鏡 "mirror" by another word for mirror, namely jian 鑑. Another example for the change of book titles is the replacement of Taixuanjing 太玄經 by Taiyuanjing 太元經 in order to avoid the personal name of the Kangxi Emperor 康熙 (r. 1661-1722) which was Xuanye 玄燁. Alternatively, the last dot of the character might be left out: 𤣥.

The most important premodern studies on the six types of characters are the treatise Liushu lüe 六書略 which is part of Zheng Qiao's 鄭樵 (1104-1162) administrative history Tongzhi 通志 from the Song period 宋 (960-1279), and the Liushugu 六書故 by the Yuan-period 元 (1279-1368) scholar Dai Tong 戴侗 (1200-1285).

Appearance and components of characters

Unlike other logographic scripts like Egyptian hieroglyphs or cuneiform scripts, the Chinese script consists of characters that all have the same size. This has to do with the fact that one character stands for one syllable and one word. A simple character like 人 occupies the same space as a complicate one like 囊. Components of characters are accordingly reduced in size:

Figure 1. Scaling of character components

The more complex characters are, the smaller are individual components. Yet standing alone, simple components like 火 or 又 occupy all the space reserved for one character.

Chinese texts can be thought of as written in a grid of square fields (fangkuai 方塊) that have to be filled to all edges. Beginners therefore use to train writing by the use of grid-pattern paper. Yet such boxed are also used for drafts.

Figures 2-3. Grid-pattern paper for writing

Left: Modern grid-pattern paper for drafting texts (written in rows from left to right). Right: Example of the manuscript edition of the Qingguoshi 清國史 (detail, written in columns from right to left). The blank cell signifies respect for the following words (xianshi 先世 "former generations"). In ready-made prints, this is done by a "line break" and positioning the first word a line higher than the usual line (see below).

Because of predefined space for each character, complex characters are often not readable in prints of minor quality. Simple characters (wen 文, modern term dutizi 獨體字) are quite rare, at least seen from the whole thesaurus of characters, while compound characters (zi 字, modern term hetizi 合體字) make out 90 per cent of all characters. Most of the latter consist of two parts, either left and right or top and bottom. One of the two parts is mostly a phonetic part (sheng 聲, the phonetic, modern term shengfu 聲符), indicating roughly the pronunciation, and the other part indicates the field of meaning (xing 形, the signific, modern term yifu 意符). This type of character, to which most characters belong, is called xingsheng zi 形聲字.

Characters can be divided into thirteen types according to their graphic composition (with examples):

Figure 4. Patterns of graphical composition of Chinese characters

The composition of characters can be divided in solitary characters (like 中), horizontally divided ones (in two or three parts), vertically divided ones (in two or three parts), and various enclosures of differnt grades, from total enclosures like 國 to enclosures from three sides or just from two sides. The last type (多) consists of two overlapping components.

The position of signific parts is relatively fixed. The components 亻彳口氵火木扌犭礻足, for instance, are mostly standing to the left, 刀卩阝攴見頁戈鳥 to the right, and 宀穴艹竹雨 to the top of other modules. This has become a standard with the creation of the chancery script, just like the sequence of the brush strokes that always go from left to right, from top to bottom, and from outside to inside.

The signific part of a character is called its radical (bushou 部首). The term bushou came up during the Later Han dynasty. The oldest surviving character dictionary, the Shuowen jiezi, classifies all Chinese characters it records into 540 radicals. During the Ming period 明 (1368-1644) the number of radicals was reduced and was fixed at 214. This number is traditionally connected with the Kangxi zidian 康熙字典 dictionary (as the so-called Kangxi radicals 康熙部首).

Character simplification in the People's Republic and attempts by various scholars to make character lookup easier has resulted in different systems of radicals. Characters that do not follow the principle of combining a phonetic with a signific element, like the huiyi or zhishi characters, are arbitrarily subsumed under one radical, for example 好 "good, to like" under 女 "wife", not 子 "child"; 去 "to erase, to do away" under 厶 and not under 土, and 不 "not" under 一, and not under 木. In the last two cases, the radicals are purely graphical and have nothing to do with the meaning of the characters.

What has also changed significantly in the last 2,000 years is the monosyllabic character of the language. Most nouns and verbs in modern Chinese are disyllabic, and it is therefore not any longer justified to say that one character represents one word. One has therefore to discern between character dictionaries (zidian 字典) and word dictionaries (cidian 辭典 or 詞典).