Introduction

The Cangjie input method 倉 頡 輸 入 法 was invented by Chu BangFu [ 朱 邦 復 Zhu BangFu ] in 1979 designed first for input of Chinese characters in Taiwan. It is based on the traditional form of Chinese characters [ 繁 體 字 fantizi ], moreover, for certain characters where more than one traditional character exists, that is there are variants [ yitizi 異 體 字 ], only one particular form is taken as the standard traditional characters. Subsequently, these are known as "computer characters" [ 電 腦 字 diannaozi ]. Its popularity has since grown, since he has placed it into the public domain. It is used for inputting of Traditional Chinese character text via a qwerty keyboard. For consistency, the method is abbreviated and called Cangjie throughout.

Cangjie KeyBoard Layout

• Cangjie is reliant on the shape of a Chinese character or hanzi 漢 字
• The Cangjie keyboard contains the twenty four mnemonics
• A mnemonic is a character which represents one or more bases.
• A base is a shape which appears within a Chinese character
• There are four sets of mnemonics as follows

  Chinese	Name	Mnemonics	Associated Key	Other Names
  哲 理 類	The Philosophical Set	日 月 金 木 水 火 土	A B C D E F G
  筆 劃 類	The Penstroke Set	竹 戈 十 大 中 一 弓	H I J K L M N	斜 點 交 叉 縱 橫 勾
  人 體 類	The Body-Parts Set	人 心 手 口	O P Q R
  字 形 類	The Character Form Set	尸 廿 山 女 田 卜	S T U V Y

• Other Names categorises the main types of Chinese character penstrokes
  • 斜 slant (/)
  • 點 dot (.)
  • 交 cross_1 (+)
  • 叉 cross_2 (x)
  • 縱 vertical (|)
  • 橫 horizontal (-)
  • 勾 hook ( 　)
• Keys X and Z are reserved, with different special usage in Cangjie

Summary of the bases for each Mnemonic

Character Shapes

The following is a non-exhaustive list of character 'descriptors' with examples.

Left and Right

明
語
川

Top and Bottom

吉
哭
晶
合

Outer and Inner

因 間 同 凶

Linked

亞 馬 事

Linked characters are those where there is one or more linear link running through a character in such a way that they do not fall into the first three categories. This linear link may not necessarily be a straight line though.

There are also portions of characters which wrap other portions as follows, rather like the Outer to Inner characters

Wrappers

凶
區
同
句
司
這

Extraction of Mnemonics

Some definitions first:

Extraction Principle

Mnemonic Extraction Rules

In order to limit the number of keys or mnemonics used to enter a Chinese character, Cangjie has a maximum of five mnemonics. Not every single character has this many keys though. The limiting of keys ensures that there is good speed of typing for experienced users.

Cangjie defines that mnemonics should be extracted from the character by splitting the character into portions, and then considering the total number of mnemonics that can be extracted first from each portion, before finaly extracting them.

• Extraction from a Linked Character
  • Linked characters extract the very first mnemonic as being the character's lead. Everything else is considered as being the character body. The character body will yield a total of three other mnemonics. This yields a maximum of only four mnemonics for a linked character as follows

    Character Lead	First Only	事	十
    Character Body	First, Second, Last		中 中 弓

• Extraction from a Partitionable Character
  • A partitionable character will have two portions which can clearly be defined as being the character lead, and its character body. A maximum of two mnemonics can be extracted from the character lead, and three from the character body. This gives a maximum of five mnemonics in total.

  • Character Leads
    • Where more than two mnemonics can be extracted from a lead, only the first and last are taken.
  • Character Body

    When there is more than one mnemonic for the body, Cangjie requires us to split the body into a secondary lead and a secondary body. The character body depends upon whether it is linked or partitionable.

    • Linked Body

      • The secondary lead of a linked body may only extract one mnemonic

      • The secondary body of a linked body may then extract two mnemonics

      • Where there are more than two mnemonics in the secondary body, only the first and last are taken from it.

    • Partitionable Body

      • Secondary Lead
        • The secondary lead may extract a maximum of two mnemonics. If there is only one, then one is taken, if there is two, two is taken, if there is more than two, then only the first and last mnemonic from the secondary lead is taken.

      • Secondary Body
        • If there are two mnemonics from the secondary lead, then only the last mnemonic is taken from the secondary body.

        • If there is only one mnemonic from the secondary lead, then a maximum of two mnemonics are taken from the secondary body.

        • Where the secondary body requires two mnemonics, and there are more than three available, then only the first and last mnemonic from the secondary body is taken.

  LEAD	One Mnemonic	Two or More Mnemonics
  	first	first and last
  	語	卜 口 are the first and last mnemonics of 言

  BODY	Type	Examples	Secondary Part	If One Lead Mnemonic	If Two or More Lead Mnemonic
  	Linked	怔 (Lead was 心 )	Secondary Lead 一	One 一
  			Secondary Body 止	First and Last 卜 一
  	Partitionable	瞵 (Lead 目 was 月 山 )	Secondary Lead 米		First and Last 火 木
  			Secondary Body 舛		Last 手
  		楂 (Lead 木 was 木 )	Secondary Lead 木	One 木
  			Secondary Body 旦	First and Last 日 一

• Characters with a dot in the top right corner Characters like the following
  求 球 犬 逑 就
  have a dot in the top right corner. This dot is considered as the highest point in the character, and it must be taken as the first mnemonic of the lead or body.

  Example	Lead	Body
  求	戈	十 水
  球	一 土	戈 十 水
  逑	卜	戈 十 水
  犬	戈	大
  就	卜 火	戈 大 山

• Conciseness Principle [精 簡 原 則 jingjian yuanze]
  When a character decomposes into many bases, the least number is correct

• Integrity Principle [ 完 整 原 則 wanzheng yuanze ]
  When equal combinations of two mnemonics can be extracted, take the more complex feature as the first mnemonic, which should leave a less complex one behind. For example, do we use 十手 or 手十 for 丰? Since the base under 手 is more complex than 十, the order is 手十

• Character Feature Principle [ 字 形 特 徵 原 則 zixing tezheng yuanze ]
  When splitting a line to extract bases, the point occurs before a turning point between the two bases. For example, when extracting mnemonics for 力 we obtain 大尸 rather than 大弓

• Omiting Parts [ 部 分 省 略 bufen shenglue ]
  These depend on the extraction principle and rules outlined above. For example, in the character 語 , the lead was 言 and only the first and last mnemonics were extracted according to the rules. The intervening two mnemonics which could have been extracted, 一 一 are omitted.

  Further, the secondary lead in the same character 五 can yield three mnemonics, but only the first and last are taken, in this case, 一 一 . The intervening mnemonic 木 is omitted in accordance with the rules.

• Wrappers and Omission [ 包 含 省 略 baohan shenglue ]
  When the last base is surrounded by three or four sides, the wrapping base is taken as the final base. For example, 喝 the final base which should be 人 , however, it is wrapped by an L shaped feature, which comes under the mnemonic 女 .

Other characters [ 列 外 字 liewaizi]

• Compound Character [ 複 合 字 fuhezi ]
  Nine special characters are singled out and only represented by two mnemonics:

  

  Compound Example	Lead	Body	Compound Example	Lead	Body	Compound Example	Lead	Body
  門	日 弓		闈	日 弓	木 一 手	憫	心	日 弓 大
  目	月 山		瞑	月 山	月 日 金	相	木	月 山
  鬼	竹 戈		魈	竹 戈	火 月	魏	竹 女	竹 戈
  鬥	中 弓		鬨	中 弓	廿 金	鬫	中 弓	一 十 大
  陳	弓 中	木 田	邦	手 十	弓 中	隰	弓 中	日 女 火
  隹	人 土		集	人 土	木	誰	卜 口	人 土
  虎	人 心	竹 山	豦	人 心	一 尸 人	處	卜 心	竹 水 弓
  幾	女 戈	竹 戈	畿	女 戈	田	機	木	女 戈 戈
  羸	卜 口	月 廿 弓	贏	卜 口	月 月 弓	瀛	水	卜 口 弓

• Special Characters [ 特 殊 字 tezhuzi ]

  These are the Chinese characters for wood 木 , fire 火 , and big 大 . When there are other character strokes which cut across these charactetrs, or overlap them, making them non-partionable, one first extracts those mnemonics which come before it, then these special characters are considered as the higher portion in the overlapping.

  • 東 木 田
  • 束 木 中
  • 末 木 十
  • 木 人 人
  • 拳 火 手 手
  • 脊 火 金 月
  • 眷 火 手 月 山
  • 陝 弓 中 大 人 人
  • 爽 大 大 大 大
  • 夾 大 人 人
  • 柬 木 田 火
  • 刺 木 月 中 弓
  • 乘 竹 木 中 心
  • 秉 竹 木 中
  • 謄 月 火 手 口
  • 賸 月 火 手 金
  • 卷 火 手 尸 山
  • 療 大 大 金 火
  • 夷 大 弓
  • 僚 人 大 金 火

  Note that the mnemonic sharing the same area overlapped by the special character 木 keeps part of the vertical, so 東 is 木 田 rather than 木 日 .

• Special Character Leads [ 特 殊 字 首 tezhu zishou ]
  Some other special character leads have be defined. These are:
  
  • 詹 弓 金 卜 一 口
  • 帝 卜 月 中 月
  • 擾 手 一 月 水
  • 產 卜 竹 竹 手 一
  • 岳 人 一 山
  • 瀝 水 一 木 一
  • 囊 十 月 口 口 女
  • 孛 十 月 弓 木

• Difficult Characters [ 難 字 nanzi ]

  There are some characters which do not fit into the above rules easily, in the fact that they have very a complicated structure, or simply that there isn't a defined base right for it. The special key X comes into play for these characters. The key X is known as 難 [nan2] or difficult.

  Difficult characters of the type :

  • [one lead mnemonic] [X] [last mnemonic in the character body] 首 難 尾
    
    • 身 竹 難 竹
      (when 身 appears 躺 射 躲 , it is a character lead, the mnemonics are 竹 竹 only)
    • 慶 戈 難 大
    • 廌 戈 難 火
    • 鹿 戈 難 心
    • 龜 弓 難 山
    • 姊 女 中 難 竹
    • 淵 水 中 難 中
    • 兼 廿 難 金
    • 黽 口 難 山

      Note that when (some of) these characters act as leads, the two mnemonic rule for character leads takes precedence.

  • [one lead mnemonic] [X] 首 難

    

    • 臼 竹 難
    • 興 竹 難 月 金
    • 叟 竹 難 中 水
    • 肅 中 難
    • 齊 卜 難
    • 卍 弓 難

Speed Cangjie

Now that you've understood the main points of Cangjie, it is possible to cut out the bulk of the work using the method called Speed Cangjie. In Chinese characters, the input method is known as 簡 易 [jianyi] or 速 成 [sucheng].

The method is simple, you take the [first mnemonic][last mnemonic] or 首 尾 and if there is more than one character, the input editor will give a list from which to select from.

The End.

• 中文輸入法使用冊 by 劉裕敏 published by 知識圖書公司

• 全圖傳意 倉頡碼字典 An Illustrated Dictionary for Changjei Code by 蔡劍南中文電腦中心 published by 萬里機構 ~ 萬里書店出 版 ISBN 962-14-1509-8

• 中文輸入法大全 published by 指南針集團有限公司 ISBN 962-8788-18-3

This page was created on Wednesday 10th April 2002, and last revised on Sunday 28th April 2002