Zhengma Chinese Input Method

[Altair]

I recently discovered the 郑码 (Zhèngmǎ) Chinese input system and have become a fan, having previously learned the basics of 五笔字形 (Wǔbǐzìxíng) and having failed to learn 仓颉 (Cāngjié) after a few abortive attempts. Since there might be one or two people with similar goals to mine I thought I might write a short commentary of Zhengma.

 

My own personal consideration in studying Zhengma is that I would prefer an input system that helps me remember how to write characters without being too cumbersome to use, learn, or maintain. I would also like one that approximates the experience of touch typing in English. That means a shape-based input system.

 

While I consider some form of pinyin input indispensable, it does not help with remembering how to write characters.  Also, one of the things I enjoy about learning Chinese is the nature of the characters themselves. Writing them out by hand can often be tedious, but skipping the process altogether feels like cheating for a learner like me and significantly decreases my enjoyment of writing in Chinese.  I have reached these conclusions, however, just as mobile devices have proliferated and made pinyin and handwriting almost the only convenient choices for those outside China.  In any case, I have found learning Zhengma simply fun.

 

What I like about Zhengma is that the vast majority of the key assignments are rule based and give much less of a feeling of arbitrariness than the other input methods I have experienced.  Since I actually do not type Chinese often, it means that I have to rely much less on mental memory or muscle memory to know what keys correspond to which character shapes.  It seems easier to maintain.

 

Zhengma and Wubizixing have a somewhat shared development history, but Zhengma uses stroke order on a much more consistent basis.  The vagaries of the long development of Hanzi do mean that there is no mechanical way to go from a character to the corresponding Zhengma code; however, in almost all cases, the code of a character has a fairly tight logical basis that almost never feels arbitrary.

 

I was initially reluctant to learn Zhengma for two reasons:  (1) I read that it was specialized for scholars and (2) the charts I read seemed to have unexplained inconsistencies in them that soured me on the system.  Having learned the system, I now see these two issues in a different light.

 

Zhengma seems to have begun as a classification system for characters.  As a result, certain parts of the system have a bias toward how you would think of characters as listed in a character dictionary or a look-up system, rather than as you would encounter them in daily typing.  In reality, it is not an issue and even has some built-in advantages.  For instance, the version of Zhengma I learned covers a very large base of characters, perhaps over 20,000 and so can handle simplified (e.g., 为) and traditional characters (e.g.,both 為 and 爲), as well as many Japanese (e.g., 竜), Korean (e.g., 畓), and rare characters (e.g., 媤).  However, the version includes shortcuts only for simplified Chinese characters and so would probably prove quite irritating to those primarily typing traditional characters or characters for languages other than Chinese.  Zhengma also relies quite heavily on codes that have multiple values distinguished by defaults weighted toward what a mainland user would expect and away from traditional characters.  Typing something other than simplified Chinese would involve a fair amount of typing numbers to disambiguate codes that would default to a simplified character.

 

Zhengma comes pre-installed in all except the latest version of Windows, but you can also use this site to type Zhengma Codes:

 

http://zhengmashurufa.51240.com

 

You can use this site to look up Zhengma codes using pinyin: http://zhengma.911cha.com/.

 

Zhengma is an input system that breaks down each character into a series of roughly 200 roots or components (字根) represented by 1- or 2-letter codes that can be classified and sub-classified in various mostly systematic ways.

 

The letters that make up the codes are generally determined according to the order and type of the first 1-3 strokes in the roots (字根) they represent, using the mainland standard of the handwritten forms (楷书) and using the stroke order of the handwritten forms.

 

Each root can first be classified into 5 stroke regions, depending on the type of its initial stroke.  There are 5 types corresponding to the first 5 strokes of the characters 木马 (Trojan Horse), in that order (一丨丿丶乙).  These 5 types correspond to the 5 regions of the alphabet.

 

The letters "A" through "H" correspond to character roots that all begin with 一 (横).  The letters "I" through "L" correspond to character roots that begin with 丨(竖).  The letters "M" through "R" correspond to character roots that begin with 丿(撇).  The letters "S" through "W" correspond to character roots that begin with 丶(点).  And The letters "X," "Y," and "Z" correspond to character roots that begin with a bent stroke such as 乛, 乙, ∠, or 乚 (折).  Roots are further divided within their respective regions according to the type of the second or third strokes.  In a few cases, the roots are distinguished according to overall shape in addition to stroke order.  Also, some letters combine two slightly different stroke orders, when they do not correspond to many roots by themselves.  Other stroke orders have been spread across two letters when they correspond to many roots.  In this latter case, there is usually some other principle, such as overall shape, that distinguishes them.

 

It is easy to remember the 5 regions by using some mnemonic to remember the first letter that starts each region.  Here is my system.

 

"A" is the beginning and so starts the roots beginning with the horizontal stroke, 一 (横). The letter "I" (as in "Ice") begins the region with roots beginning with the vertical stroke丨(竖) and can be remembered by noting that the shape of the letter "I" is virtually identical to a vertical stroke.  "M" begins the region with roots beginning with 丿(撇).  I remembered it with some simple mnemonic like "'M' for move to the side."  Another option is to think of "me" for 我, which begins with 丿(撇), and is the high-frequency character associated with "M."  The next is "S," which begins the region corresponding roots beginning with a dot.  You can remember it by associating it with "spot" or the dot that begins the handwritten version of 言  ("say"). The last letter starting a region is "X," which is for roots beginning with a bent stroke 乙 (折).  I remembered it by its inclusion in the phrase "XYZ" or by its shape of 4 bends.  A mnemonic for all 5 could be "AIMS for the X."

 

To learn the complete basics of Zhengma, you can consult these sites:

 

http://www.scribd.com/mobile/doc/6082054 (in English, includes setup instructions)(The explanation is a good introduction, but omits some information that I found important for my learning style.)

 

 

http://wenku.baidu.com/view/7f0180e3524de518964b7dd8 (thorough lessons, but in Chinese)

 

 

To get the basics, you need to begin with the 50 or so First and Second Principal Roots.  Using on-the-spot mnemonics, I found this easy to do.   These Principal Roots correspond rougly to two roots for each of the 26 letters of the alphabet that mostly have a shared stroke order for the first 1-3 strokes.  The First Principal Roots were chosen mostly for frequency and so have some inconsistencies in how their actual stroke order matches the ideal, but the Second Principal Roots are consistently illustrative of the required stroke logic.  Once you learn these, you can learn to recognize the Secondary Roots, which are mostly combinations of the Principal Roots or the stroke orders they represent. You don't need to memorize all these Secondary Roots to be able to type, but need familiarity in order to make educated guesses as to what is a root and what isn't and what the codes are likely to be.  You can find a good chart at either of the following sites:

 

http://imgsrc.baidu.com/baike/pic/item/024f78f0f736afc3027f763db219ebc4b7451272.jpg. (Study the region in blue first.)

 

http://hi.baidu.com/ycckkk2008/item/606467dcd978f348ddf9bef8 (Study the last column first)

 

If there is interest, I can also add my own clarifications or even a beginning tutorial.

 

 

 

[skylee]

Why do you prefer Zhengma to a pure stroke order input? (I don't know of a stroke order input for use on PC besides 九方, but it is very commonly available on mobile phones.)  I don't understand why it is a good thing to remember those roots.  I think it is a good thing to be able to input characters by either just knowing how to write them (in stroke order, because that is how characters are written) or pronounce them.  Any other requirements do not seem necessary.

[Altair]

I think you have said before that you prefer input systems that don't involve much additional learning. I can certainly understand that view. But one of my purposes in learning the system was to learn a way to type Chinese that approximates touch typing in English, without feeling alien to the nature of the characters. I want to feel that I am typing Chinese rather than English, without actually have to write it out stroke by stroke.

I have almost no experience inputting characters using strokes, but I am pretty sure it would feel just as tedious to me as writing them out by hand. Just as I find looking up characters by stroke-count tedious.

I realize that stroke input is a popular method for cell phones, but I hate texting even in English, since using two thumbs is not the same as touch typing with nine fingers.

Consider inputting the first verse of the Romance of the Three Kingdoms:

滚滚长江东逝水

I don't know whether predictive text can handle that verse or not, but the phrase has about 54 strokes spread over only 5-10 stroke types. That means a lot of strokes concentrated on a few keys and a few fingers. Typing it with Zhengma takes about 21 letters spread almost over the entire 26-letter keyboard. I especially hate typing that involves a lot of repeat movements, because of the finger fatigue it often causes.

I can understand why memorizing 200 roots seems totally unnecessary; however, with the Zhengma system you really don't need to memorize them all, since so many can be guessed. Knowing what I now know, I could have taught myself the basics in 1-2 hours. People who are new to the system, who are less geeky than me, and who have no one to explain things in a simple way would need more time, but still much less time than I think it takes to learn Cangjie.

Basing my inputing on strokes alone would feel like printing stroke by stroke in English, which I also hate to do beyond printing a few words here and there on forms. Basing my inputting on the 200 or so Zhengma roots makes me feel that I am just typing character components, which is how I mostly think of characters anyway. It also comes in handy since the little typing I do is often of things that predictive input systems are not good at. Now if I only my ability to produce characters were anywhere near my ability to recognize them, I would be set.

[skylee]

After reading your first post I searched about Zhengma, but as you already know I am impatient about such input methods.  Before long I found something else I was interested in.  It is an ime called Rimeime.  I have used google pinyin input for a long time and was content, kind of. The only major shortcoming of google pinyin is that it is based on simplified characters and because of this, even though I have set the output to traditional, it gives me a simplified font in MS Word.  It is not really that "major" a flaw, but I don't particularly enjoy it.  Now I have removed google pinyin and am using Rimeime only.  So far so good.  Rimeime has different versions for linux, windows XP and 7 and mac, but there is not yet a version for windows 8. I think there might be some bugs in this ime's menu, but I am going to use it at least for the time being.

 

Well I have found something I like, and you have found something you like. 

[Gareth422]

Sorry to necro this thread, but this is currently the most in-depth explanation of Zhengma in English, anywhere on the web (at least as far as I can tell). I'd be very interested if you wrote a tutorial, like you suggested.

[Gareth422]

Oh, and I gathered together some information about how to get Zhengma on all sorts of different platforms:

 

Windows:

1) Download Rime at rime.im, and install it.

2) Download the files for Zhengma here.

3) Follow these instructions to modify the configuration files to add it to the list.

4) Run WeaselDeployer.exe and enable Zhengma.

 

Linux:

 

1) Install IBus from your distribution's repositories, along with Zhengma.

2) That's pretty much it.

 

Android:

1) Install Baidu's input method.

2) Download the Zhengma table here , then rename it to have a .def extension.

3) Add it as a custom table.

 

Instructions (in Chinese) are available here: http://www.augsky.com/860.html

[NinKenDo]

I'm curious. How do I go about attaining the knowledge necessary to know which key leads to which characters? This is still really opaque to me even after OP's great explanation.

[Gareth422]

I've just started to develop an intuitive feel for it. I can't tell you anything that isn't explained in any of the links above, reality. Three guide on Scribd is the first place to look. After that, keep the chart handy. When you can't figure out how to type a character, look up the code, and try to understand why it has that code. Eventually, c you'll figure out the patterns and quirks of it.

[NinKenDo]

Hey guys, I tracked down another great resource for ZhengMa which I think, in combination with OP's post, makes ZhengMa actually very easy to understand. Thanks to OP and the author of this document (Scribd PDF) [or HTML version here], I feel really confident about learning ZhengMa now.

[softwood]

ZhengMa IME is a brilliant input method! I strongly recommend it!

I also learned it after struggling with WuBi... I can compare them: WuBi is very weak compared to ZhengMa.