Take the 2-minute tour ×
Chinese Language Stack Exchange is a question and answer site for students, teachers, and linguists wanting to discuss the finer points of the Chinese language. It's 100% free, no registration required.

I would like to be able to compute the distance between 2 hanzi according to their components.

My idea is that a hanzi like 式 is made of 工 and 弋. 2 hanzi are close if they share common components.

To be able to compute such distance, I need to be able to have a list of hanzi and their components in a digital format. Does such list exist?

share|improve this question
    
I don't know what your purpose of doing this is. If you want to improve your Chinese language proficiency, then you don't need to take this approach, because all of Chinese native speakers were not learning Chinese in this way. If you want to do some academic research on how Chinese characters are constructed and how inter-connected to one another, then you may try this method, but I doubt it works. The reason is that Chinese characters have been evolving for thousands of years, and the face of characters got changed drastically. It is very difficult, if not impossible, to figure out the origi –  孤影萍踪 Mar 26 '13 at 22:02
    
there is obviously, as in every language and script, going to be a difference what is synchronously and historically close or related. a given element may be shared by two characters today but have different origins and vice versa. that said, your argument is faulty—just bc, say, english speakers do not make lists of common morphemes when learning the language as a child doesn't mean that's meaningless for a L2 learner (or not interesting for a native speaker). –  flow Dec 16 '13 at 12:26
add comment

4 Answers 4

up vote 11 down vote accepted

I know exactly the thing. There's a project on Wikimedia commons to document the substructure of characters in terms of other characters. I haven't spent much time with the data itself, so I can't tell you how complete it is, but it seems pretty good based on my experience with the Tatoeba character search tool that is based on the data. That tool allows you to "explode" characters into their components (e.g., 蝴 => 月虫...) or search for all characters with a particular component (e.g., 隹 => 倠傕傩...).

Another project to do something similar is Wenlin's database of characters in the Character Description Language. Here is Wenlin's page of resources on the CDL. I'm not really sure if you have to buy Wenlin in order to get at some data. Anyhow, it looks pretty awesome.

Best of luck in your endeavors! I've wanted to do something similar, but I've always been too busy. Make sure to tell us about the cool things you do with the data :-)

share|improve this answer
    
Do you work on wiki projects? –  deutschZuid Mar 27 '13 at 10:05
    
@deutschZuid No I don't. But I'm a computery guy, so Chinese-related datasets are pretty darn interesting to me. –  Stumpy Joe Pete Mar 27 '13 at 14:49
    
I am actually developing tools for Tatoeba haha. –  qdii Mar 27 '13 at 16:09
    
@qdii Hahaha, well your existing tool seems to have the data you need. –  Stumpy Joe Pete Mar 27 '13 at 16:42
add comment

Check out hanzicraft.com & zhongwen.com; they have functions that seem similar to what you're asking for.

share|improve this answer
add comment

You can looking for the hanzi in this page http://www.mdbg.net. Then search the hanzi, when you found it click it. Then you can click over the (1) scissors and show it!

exemple: https://lh6.googleusercontent.com/-0xcJU0i3rns/UWB7C5ji4SI/AAAAAAAAA2Y/6P1xzl-fkBo/w570-h428-p-o/Selecci%25C3%25B3_001.png

(2) After you can click over the components!

share|improve this answer
add comment

You can use Gavin Grover's CJK Decomposition data (used by HanziJS and cjklib).

share|improve this answer
    
+1 Awesome resources I haven't seen before. –  Stumpy Joe Pete Mar 28 '13 at 2:15
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.