Chinese Kanji Etymology Reference Terms
– Free Online Kanji Etymology Dictionary –

S/B: Shell and Bone Characters
The earliest forms of the characters are those carved into shells and bones (chia ku wen: 甲骨文; in Japan, more commonly known as 甲骨文字) and those (jin wen: 金文) inscribed into tripod kettles, tools, weapons, musical instruments and other objects made of bronze.

The shells and bones were used in divination rites during the Shang Dynasty (1600-1066 B.C.E.). Heat was applied to induce cracks in the shells and bones, and fortunes were cast based on the patterns of the cracks. Afterwards, details concerning the provenance of the shells and bones, the diviner, the question posed for divination and the interpretation made were inscribed with bronze knives or the edges of hard jewels, creating characters thin and angular in form.

It is worth noting that fully 5,000 distinct characters, some of a high degree of morphological development, have been identified among the shell and bone characters. This suggests that the characters will ultimately be found to date much further back than can be confirmed at present.

Return to Top | Return to Previous Page

BIS: Bronzeware Inscription Style
The bronzeware characters, which record legal affairs, proclamations, or the performance of sacred rituals, commemorate achievements and so on, these feature bolder lines and are more highly ornamented than the shell and bone characters. The forms of the characters remained fairly static during the first centuries of the Zhou Dynasty (1066-256 B.C.), but by the time of the Warring States period (475-221 B.C.), the collapse of central authority had created a profusion of variant styles.

Return to Top | Return to Previous Page

SIS: Seal Inscription Style
Shi huang di (始皇帝), founder of the Qin Dynasty (which can be dated from either 246 or 221 B.C. - 206 B.C.), made a priority of rationalizing the writing system. He favored hsiao chuan (小篆), a curvy style used in carving seals.

Return to Top | Return to Previous Page

SCR: Scribal Form
During the Han Dynasty (206 B.C.E - C.E. 225, excepting an interregnum from C.E. 8 - C.E. 25), a more angular style with clean, straight lines came to be preferred by scribes who spent their lives writing documents. Known as k' ai shu (楷書), this is the style that serves as the model for the printed characters used today.

Return to Top | Return to Previous Page

ARC: Archaic Style
A general term for a variety of pre-Scribal Form styles.

Return to Top | Return to Previous Page

Types of Characters
The Shuowen Jiezi (説文解字), a famous character dictionary compiled 100 C.E. by Xu Shen (許慎), identifies six types of characters. However, two of these six types have to do with the usage rather than compositional structure. The types with which we concern ourselves are pictographs (Chinese: 象形字; Japanese 象形文字), ideographs (Chinese: 指事字; Japanese: 指示文字) and compound characters, the last group of which are traditionally considered to be of two varieties: semantic compounds (Chinese: 會意字; Japanese: 会意文字) and phonetic compounds (Chinese: 形聲字; Japanese: 形声文字).

Of the tens of thousands of characters, pictographs number in the mere hundreds, but as elements in compounds they are of the greatest importance. Examples of frequently used pictographs include 刀 sword, 口 mouth, 女 woman, 山 mountain, 心 heart, 手 hand, 日 sun/day, 月 moon, 木 tree, 水 water, 火 fire, 田 field, 目 eye, 石 stone, 竹 bamboo, 虫 insect/creature, 足 leg, 車 vehicle, 門 gate, 雨 rain/natural phenomena, 馬 horse, 魚 fish, and 鳥 bird.

Ideographs are representations of abstract concepts rather than pictographs of objects. They are few in number, but include key characters as下 under/below and 上 over/above, as well as　一 one, 二 two, 三 three, 五 five, 七 seven, 八 eight, and 十 ten.

Compound characters are composed of two (rarely more) elements. Traditionally, compound characters were said to be of two types, phonosemantic compounds and semantic compounds. However, the second category, semantic compounds, was alien to the earliest inventors and users of the characters, as all compound characters in the early oracle bone and bronzeware inscription styles are phonosemantic in nature.

For the record, phonosemantic compound characters contain one element (a sound note) suggesting the character's pronunciation and meaning, along with a second, semantic element hinting at the character's meaning alone.

In contrast, semantic compounds have been said to contain no sound note, but rather to consist of two or more elements functioning purely semantically.

Why did certain compound characters, some 250 or so, come to be regarded as semantic compounds?

In short, because the sound note element in these characters became unrecognizable on account of 1) Consonant shift in the initial or the final 2) Use of variant or abbreviated forms 3) Transformation in graphic form of the sound note element, or 4) Borrowed pronunciations.

There is no authoritative compendium of semantic compounds, but reproduced here or here is the list of basic characters taught to Japanese elementary school children, with semantic compounds segregated from the other types of characters. Looking up these characters in either the Kanji Networks database or on the Chinese Characters Concept Map will reveal the particular reason, among the four noted above, why those of them attested in the oracle bone and bronzeware inscription styles are in fact phonosemantic compound characters.

Phonetic compounds account for the vast majority of Chinese characters. Here, as noted above, one element in the compound (the sound note) is intended to suggest the character's pronunciation. In this dictionary, the phonetic element is denoted by (Type [#] phonetic).

Phonetic compound characters are of eight types.

　
• Type 1: The phonetic element and the character itself belong to the same word family. Example: 夬 KUAT and 決 KUAT.
　
• Type 2: The phonetic element and the character itself differ in the final consonant. Example: 胥 SAG and 婿 SAR.
　
• Type 3: The phonetic element and the character itself differ in the initial consonant. Example: 甚 TAM and 勘 KAM.
　
• Type 4: The phonetic element and the character itself differ in the vowel. Examples: 曼 MUAN and 慢 MAN; 皮 PAR and 坡 PUAR; 而 NAG and 需 NUG; 宗 SUNG and 綜 SONG; 充 TONG and 銃 TUNG.
　
• Type 5: The phonetic element differs from the character itself in two respects, yet is nevertheless properly regarded as the sound note. Examples: 卒 SUAR and 萃 TUAT; 乃 NAG and 秀 SOG; 無 MAG and 撫 PUAG; 光 KUANG and 恍 MANG.
　
• Type 6: The phonetic element functions onomatopoeically or mimetically, and not for its customary meaning. Examples: 瓜 simulates the sound of a crying baby in 孤; 立 simulates the sound of rushing wind in 颯; 此 simulates a hissing noise in 呰 and so on.
　
• Type 7: The phonetic element transliterates the sound of a term non-Han Chinese in origin. In this case, the phonetic element may or may not have a secondary, semantic function. Examples: 革 in 靴; 留 in 瑠; 匍 in 萄 and so on.
　
• Type 8: The phonetic element is an abbreviated form of a more complex character. Examples: 舌 in place of 括 in 蛞; 各 in place of 洛 in 路; 刃 in place of 忍 in 靱 etc.

Consonants and Consonant Networks
The structure of ancient Chinese was extremely simple: an initial consonant (also referred to simply as an "initial") + a final, the latter composed of a nuclear vowel and a coda consonant. Sometimes included in the final was a medial glide (I or U).

The initials were K, L, M, N, P, S/TS, and T. Each consonant expressed an identifiable concept (see below). Word families of terms beginning with the same initial represent branches and sub-branches of the particular consonant network.

The S/TS consonant network is actually a sub-network of the T group, as we can determine by comparing word family pairs such as TAM/SAM, TAN/SAN, TAR/SAR and so on. A careful examination of these pairs reveals that the S/TS terms nearly always represent the same theme as the corresponding T terms, but with specific reference to objects that are small/thin/slender. Aside from the distinction of smallness, S/TS network terms often express the idea of irregularity in shape, whereas T network terms are normally concerned with regularity. In this respect, compare the TAR and SAR word families.

Here are the seven consonant networks, with their primordial concepts and (where applicable) variant initials.

Vowels
The main vowel in proto-Chinese was short A, or something between short A and short E. This sound linked the initial and final consonants, and exercised no semantic influence on terms. Terms with this vowel account for approximately 3/4 of the proto-Chinese vocabulary.

The other nuclear vowels were O and U; these functioned semantically, with significations both similar and overlapping. The O vowel suggests CURVATURE, sometimes extending to CURVE AND SURROUND/ENVELOP. The U vowel, when it is the lone vowel in a term ( KUG/ KUK, SUG/ SUK etc.) refers to a CIRCLE (whether full or nearly full) or to a bulky/lumpy MASS. However, when the U vowel is employed as part of the diphthong UA ( KUAN/ KUAR/ KUAT, SUAN/ SUAR/ SUAT etc.), the signification is the O vowel sense of CURVING/ROUND. The reason for using U in the dipthong rather than O was probably a matter of economy in pronunciation: voicing the sound KOAN distinctly, for example, demands more effort from the jaw muscles than does the sound KUAN.

Return to Top | Return to Previous Page

Finals
Complementing the seven initials are seven finals. Here they are, with their primordial concepts and (where applicable) variant forms.

Consonant Clusters
As shown by the NG final, consonants were sometimes clustered. A small number of initials demonstrate clustering involving the consonant L. These are KL (as in KLAM and KLAP), HL (as in HLAG), GL (as in GLOK), and ML (as in MLAK or MLAG). In each initial cluster, the separate consonants exert their semantic influence on the terms. For example, KL terms refer both to a frame and to a continuum, ML terms to concealment and a continuum, and so on.

Return to Top | Return to Previous Page

Word Formation Mechanism
As noted above, the initial conveyed a term's general sense. The presence or absence of a medial glide adjusted or delimited the range of the term. The final gave the term its specific, concrete meaning. Click here for examples.

Return to Top | Return to Previous Page

Research Background
19th century linguists such as Joseph-Marie Callery (Systema Phoneticum Scripturae Sinicae: 1841) gave impetus to the scholarly reconstruction of the sounds of ancient Chinese. In Word Families in Chinese, a landmark article that appeared in the Bulletin of the Museum of Far East Antiquities in 1934, the Swedish Sinologist Bernhard Karlgren offered a systematic method for denoting the relations between phonetically close terms.

Later (in 漢字語源辞典 [1965] 東京：學燈社, and 漢和大辞典 [1978]　東京：学習研究社), University of Tokyo professor Akiyasu Todo augmented Karlgren's work by classifying word families in Chinese into seven distinct groups, based on the initial consonant by which Chinese terms were originally pronounced. These seven groups correspond to the initial consonant networks described above, with the exception that Todo labeled the S group "TS."

As outlined above, the Kanji Networks materials expand on Todo's contribution in four ways.

First, by noting the existence of a distinct semantic function in each of the seven consonant groups identified by Todo.

Second, by noting that not only the initial consonant but also the final consonant in ancient Chinese exerted semantic influence.

Third, by noting that the vowels O and U also exerted semantic influence on terms in ancient Chinese, suggesting curvature, roundness or massiveness.

Fourth, by delineating the eight distinct types of the sound note occuring in phonetic compound characters.

Return to Top | Return to Previous Page

Points of Note Regarding the Characters
Previous Forms
Until the Second World War, the forms of the Chinese characters employed in the various Asian countries were quite similar. Toward the end of the 1940s, however, Japan simplified hundreds of characters, and China took an even more dramatic simplification path in the 1950s. Here, we concentrate on the previous and present forms of Japanese characters.

Examples of old and new styles include the following. Present styles precede the former.

Alternate Forms
Japanese and English-language kanji reference sources distinguish between the following forms of characters: standard (本字), traditional (旧字), non-standard (別体字 or the more inclusive 異体字) popular, non-authorized (俗字) and abbreviated (略字) forms. (The latter two also fall under the 異体字 rubric.) This applies both to Everyday Use Characters (EUC) and all others outside the list. It should be noted that use of the term non-authorized begs serious questions concerning who is empowered to do the authorizing, as well as the basis for determining which of alternate forms get noted and which do not.

Again, with respect to the non-EUCs, forms offered as standard can vary from publisher to publisher, and may be in turn occasionally at odds with what can be produced with the JIS character codes (see next paragraph). The most notable example in this regard would be the くさかんむり element; publishers tend to go with the four-stroke ┥┝, JIS with the three-stroke.

Part of the confusion owes to inconsistencies in the JIS (Japanese Industrial Standard) character codes. These codes, which were determined by the Japanese Standards Association, a group under the aegis of the Ministry of Economy, Trade and Industry, feature inconsistencies of the following sorts.

Borrowed Meanings
Sometimes, one or more meanings of a given character have nothing to do with its etymology. This is because the character was "borrowed" to act as the written form of another term, normally a homonym. The borrowed meaning was attached to the character, as it were, with no regard for the character's original sense. (An example of borrowing in English would be "gay," a word that entered English in the 13th century and for hundreds of years meant "happy," "merry" or "bright" before acquiring its present sense during the 20th century.)

Returning to kanji, the borrowing might take place in China, or in later in Japan. In some cases, the borrowing was prompted by graphic similarities between characters. One example from China involves the character 考, with its borrowed meanings of "think" and "consider." These meanings were taken from a separate character, 攷. 考 originally indicated simply an elderly person with a bent back (compare 老), but owing to graphic similarity with 攷 ("investigate deeply") and the fact they are homonyms (both anciently pronounced KOG), the meanings "think" and "consider" came to be attached to 考.

Expediency was often the reason behind borrowings. For instance, 御 has a minor meaning of "defend," which properly belongs to the more complex character 禦. What happened was that one or more ancient authorities decided to omit the 示 element at the bottom, thereby saving the trouble of writing five strokes. Of course, they could only afford to do this because they were confident their readers would know enough to attach the sense of "defend" to the character 御. This usage eventually gained general currency, and so the character 御 came to merge with the meaning "defend."

Borrowings in Japan include the meaning "shellfish" in 介 (on account of the resemblance between 介 and 貝) and the sense of "unable to do" in 兼 (originally, "take an object in hand").

Return to Top | Return to Previous Page

Atavistic Meanings
Some compound characters have what may be called atavistic meanings. These meanings are drawn from a constituent element of the character in question rather than from the character itself. Examples include "control" in 管 (from 官), "hang," "droop" and "suspend" in 垂 (from a combination of 土 earth and a pendent ear of grain), and "palace" in 殿 (from the left-hand element).

Return to Top | Return to Previous Page

Reverse-Image Graphic Forms
As shown by their ancient forms, certain pairs of characters are reverse images of each other. Among the everyday use characters in Japan, this applies to 片 and 爿, 育 and 子, 幻 and 予, 抑 and 印, 后 and 司, as well as 乏 and 正.

Return to Top | Return to Previous Page

Replacement Characters
In the long course of Chinese history, many characters lost their original meanings and came to refer to something completely different (see the section on Borrowings). When necessary, the Chinese devised replacement characters by tacking on a radical. Often, 人 person serves this function, as in 傾, 佐, 債, 償, 偵, and 伸 (the replacements for 頃, 左, 責, 賞, 貞, and 申). Other examples include 征 for 正 (adding 彳 movement), 蓄 for 畜 (adding 艸 grass/plant), 枝 for 支 (adding 木 tree/wood), 搬 for 般 (adding 手 hand), 燃 for 然 (adding 火 fire), 裘 for 求 (adding 衣 cloth); 蜴 for 易 (adding 虫 insect/creature), and 語 for 吾 (adding 言 words).

Return to Top | Return to Previous Page

Shape Indicators
It is common for radicals to be employed not for the meanings they normally convey but for senses suggested by their shapes. This is particularly true for 厂 (normally "cliff"): its convenient right-angle shape is pressed into service for meanings as diverse as a hanging cloth or pelt, a cover, a roof, furrows in a field, and a cooking stove. To take a few other examples from the everyday use characters, we find 艸 (grass/plant) representing horns, 火 (fire) suggesting a symmetrical object, 玉 (jewel) standing in as a musical instrument, 舟 (boat/ship) indicating the contours of footgear, and 彡 (pattern) imitating fronds.

Return to Top | Return to Previous Page

Characters Reinterpreted in Japan
Dozens of characters have been interpreted one way in China, and a different way in Japan. For example:
錠 (In China): As per 定 (fixity) + 金 metal → solder (metallic compound that binds objects together) → ingot; (medicinal) tablet (← objects of prescribed or fixed shape). 錠 (In Japan): As per 定 (fixity) + 金 metal → metallic device that fixes something in place → lock.
谺 (In China): 牙 fang/mesh (→ cross) + 谷 valley/gorge → zigzagging gorge (i.e., with a course that repeatedly crosses back and forth over itself). 谺 (In Japan): 牙 fang/mesh (→ cross) + 谷 valley/gorge → echo (← sound that crosses a valley).
鐚 (In China): As per 惡, the previous form of 悪 (be pressed upon) + 金 metal → the skirt of a helmet (which is pressed upon the wearer's neck by the helmet). 鐚 (In Japan): 惡 bad + 金 money → bad money.

Return to Top | Return to Previous Page

Repositioning of Elements
The same elements are sometimes employed in two characters. In duals such as 讎 and 讐, 裏 and 裡, 闊 and 濶, or 蟹 and 蠏, the respective characters are simply variants of each other. In other duals, however, such as 忙 and 忘, 含 and 吟, 棘 and 棗, 鵞 and 鵝, 枷 and 架, 紊 and 紋, 稾 and 稿, or 脅 and 脇, the separate characters bear independent meanings.

Return to Top | Return to Previous Page

Tripling or Quadrupling of Elements
A number of characters contain the same element tripled. The tripling suggests the idea of "many." Some examples include 三, 品, 森, 晶, along with the right-hand element in 燥 or the 劦 element in 協. Also, the previous forms of a number of characters contain tripled elements: 參 (now 参), the top element of 壘 (now 塁), 聶 in 攝 (now 摂) and so on. Note that 三, 森, and 參/参 all belong to the same word family, the concept of which is "Collect a number of items."

Rarely, an element is quadrupled within a character. 器 is one such example, as is 斷 (the previous form of 断). Quadrupling 龍 dragon by aligning two on top and two on bottom produces a 64 stroke character offered in some dictionaries with the meaning "verbose."

Return to Top | Return to Previous Page

Transformation of Elements
Elements sometimes morph into other, graphically similar but originally distinct elements. This process has occurred both in China and Japan.

For instance, the 几 elements in 机 and 冗 are now identical in appearance. However, the 几 in 机 was originally a pictograph of a pedestal, while the 几 in 冗 was originally an ideograph showing two curving lines serving as an abstract representation of the idea of "sloppy/loose." As for 宿, this character appears to contain the element 百 (one hundred), but an early form of 宿 reveals the element to have been two people sleeping with their knees pulled to their chests.

In Japan, distinctions between certain characters have been muddied by post-war alterations of standard forms (see above). The 免 and 東 elements in 逸 and 練 are two examples, having once been the similar but etymologically distinct 兔 and 柬. Compare also the previous form of 盗, which was 盜.

Of course, it also happens that certain elements transform into other, quite different elements. Whether the transformations were deliberate or the results of miscopying is a matter of interpretation. Examples from China include the 月 element in 服, 前, 朕, and 勝 inter alia (originally 舟), the 示 element in 票 (originally 火), the 言 element in 設 (originally a pictograph of a chisel), the 辛 element in 辞 (originally 司) and so on.

On the other hand, some elements that now appear independent were once more familiar in form. Examples include 急 (originally 心 + 及), and 春 (originally 日 + an element combining 屯 [shoot swelling up below the surface of the earth] and 艸 grass/plant).

Return to Top | Return to Previous Page

Abbreviated Forms
Often, an element appearing within a character serves as an abbreviated form of a graphically similar element. We can see this tendency at work in characters such as 泣, 殖, 皇 and 胎, where 立, 直, 王, and 台 act as abbreviated forms of 粒, 植, 自 and 始, respectively. As stated above, for the Chinese, the important thing was that an element suggest the proper sound; graphic considerations were secondary.

Return to Top | Return to Previous Page

Action Indicators
Five elements having something to do with a hand account for nearly all action indicators. These are 手, 又, 寸, 殳 (originally, a hand placing an object upright) and 攴 (variant form: 攵), originally a hand wielding a stick.

The only other action indicator used with any frequency is 彳. Drawn from the left side of 行 go/crossroads, 彳 normally indicates "movement," but functions as an action indicator in characters such as 律, 得, 徳, 徹 and 循.

Among other elements employed as action indicators are 廾 pair of hands (in 弊), 夂 leg (in 腹), and 辵 movement (in 選).

Return to Top | Return to Previous Page

ON and KUN Readings of Kanji
In Japan, kanji are pronounced according to what are known as ON and KUN readings. ON (音) readings are Japanese emulations of the sounds of the characters as they were pronounced on the Asian mainland at the time they were imported to Japan. KUN (訓) readings are native Japanese terms that were assigned to kanji of like meanings. For example, the character 芋 (potato) has an ON reading of ウ and a KUN reading of いも. All characters created in China have ON readings, though in modern Japanese these readings are sometimes obsolete, such as with 皿, 貝 and 箱. Meanwhile, not all characters bear KUN readings, but those that do may have multiple readings. For example, 生 and 下 each have ten KUN readings in common usage.

The background of the ON readings is quite complex. Beginners can get by if they bear in mind that there are two main types of ON readings, explaining why 下 is pronounced KA in 地下 but GE in 下品, or why 間 is pronounced KAN in 間隔 but KEN in 世間. The following explanation is for those with a deeper than average interest in the development of Japanese.

There are four types of ON readings: GO (呉), KAN (漢), TOSO (唐宋) and Customary.

GO readings take their name from the 呉 Wu (Japanese: ゴ) Kingdom (C.E. 222-280). The Wu Kingdom was one of the trio of entities that comprise the Three Kingdoms Period (C.E. 220-280) of Chinese history, the others being the 魏 Wei (220-265) and the 蜀 Shu (221-263) Kingdoms.

The GO readings penetrated Japan from the Asuka (飛鳥) Period (592-645 or 592-710) through the Nara (奈良) Period (710-784). GO readings are of two types. The first are readings taken from the spoken language of the Yangzijiang (揚子江) Basin as filtered through the Korean kingdom of Paekche (百済). The second readings were transmitted directly to Japan from Zhejiang (浙江省) Province, where the 呉 Wu Kingdom had been sited.

GO readings that remain in use in contemporary Japanese are often connected with Buddhist terms, though they are the sole ON readings for common characters such as:

応（オウ not ヨウ）；教（キョウ not コウ）；限（ゲン not カン）；七（シチ not シツ）；妻（サイ not セイ）；晶（ショウ not セイ）；浄（ジョウ not セイ）；礎（ソ not ショ）；俗（ゾク not ショク）；庁（チョウ not テイ）；痛（ツウ not トウ）；難（ナン not ダン）；忍（ニン not ジン）；能（ノウ not ドウ）；百（ヒャク not ハク）；普（フ not ホ）；便（ベン not ヘン）；毎（マイ not バイ）；脈（ミャク not バク）；民（ミン not ビン）；務（ム not ブ）；眠（ミン not ベン）；免（メン not ベン）；黙（モク not ボク）

KAN readings are those of the Sui (隋 A.D. 581-618) and Tang (唐 618-907) Dynasties. Some scholars believe that Sui Dynasty readings were making their way into Japan simultaneously with the GO readings, but generally the KAN readings are considered to have been transmitted during the early Heian (平安) Period (794-1192).

The KAN readings are based on the language spoken in the capital of the Sui and the Tang, Changan (長安: present-day Xian [西安] in Shaanxi [陜西] Province). The pronunciations were brought back to Japan by visiting emissaries and students.

TOSO readings are those of the Sung (宋 960-1279 A.D), Mongol (元 1271-1368), Ming (明 1368-1644) and Manchu (清 1616-1912) Periods. They were brought into Japan by merchants and Buddhist priests in the Kamakura (鎌倉) Period (1192-1333) and on down. Examples of TOSO readings include 杏子（アンズ）, 行脚（アンギャ）, 椅子（イス）, 提燈 or 提灯（チョウチン）, and 蒲団（フトン）.

Customary readings are those that were transmitted imperfectly. Examples include 重（ジュウ；KAN チュウ and GO ジュ）；験（ケン；KAN and GO ゲン）；崇（スウ；KAN シュウ and GO シュ）；軍（グン； KAN and GO クン）

In extreme cases, a character can be equipped with all four types of ON readings. One such character is 杏 (apricot), the ON readings of which are ギョウ (GO); コウ (KAN); アン (TOSO); and キョウ (Customary).

However, most characters have only one or two readings in active use.

Return to Top | Return to Previous Page

Kanji: The Elemental Logic of Chinese Characters
Ebook with concise etymologies of all 2136 Joyo Kanji. An invaluable resource.

On Sale Now »