Meaningfully learning Chinese characters: network based approach and original meaning based learning

We are a group researchers on complex network, working on currently on the structure of Chinese characters. We are mostly from physics, mathematics and natural sciences backgrounds. In this work, we start with the structural decomposition of each Chinese character, and represent all the characters as a network of Chinese characters or a map of the Chinese character. We then study the importance of each character in different senses, and the issue of the optimal learning order of Chinese characters. We have made some progress, though research is still in progress. The website is designed to use the research outcomes of the teams to promote the learning of Chinese and Chinese characters.

The relations of the structural components of Chinese characters to teaching Chinese have been noticed by many scholars and educators, for example, Wang Ning, Lu Bisong, Zhang Peng Peng, Joël Bellassen. Understanding the subcomponent structure of a character, and the subcomponent meanings is a good way of understanding of the original meaning of the Chinese character. It also helps to understand how the character acts within a multi-character word, and acquisition of other related words, and even sentences. For example, "木wood", "林forest", "森forest", understanding one character allows us to understand 3 characters, the learning cost is greatly reduced. Our basic questions are: First, can we systematize all characters in this fashion? Second, if the substructure at the character level is common, then can we find an optimal learning sequence of all most used characters from such a connected map of characters? For the first question, we need to extract them in etymology and relate original meaning of all the characters to its etymology, and for the second question, we need to perform network analysis on the map of characters.

In this network of Chinese characters, each node is a character. If one character use another character as a component, then a directed edge from the component to the compound character is drown to indicate such a component/compound relation. For example, from the "man 大" and "a line 一 " we get "heaven 天". We will do a hierarchical structure decomposition. For example, from "bright 照" on the first decomposition we get "火 fire " and "clear 昭" and after another step of further decompistion we get "sun 日" and "call 召" and on the third decomposition we get "knife 刀" and "mouth 口", however not any further. Our decompositions of the Chinese characters stops at basic ideographic or phonetic units, but not strokes.

You may also download a PDF format of the full map of simplified Chinese characters or traditional Chinese characters.

Of course there are possibly mistakes in our decompistion of Chinese characters, and might also in our explanations of original meanings of Chinese character. In fact, that is one motivation that we provide all of all data for the general public. We have designed a special interactive system for registered users, where users can help us to improve our decompistions and original meanings, even participate the studies by providing us better illustrative drawings and photoes for explaining the original meanings of the characters. On all those and related issues, we hope we can get our users' support and love.

Based on our own research on these issues, as well as information collected by others, on the initial construction of the site, we provide the following resources:


  1. 说文解字
  2. 文字学概要(裘锡圭,商务印书馆,1988)
  3. 常用字解( 白川静,九州出版社)
  4. 古文字诂林(李圃 主编,上海教育出版社)
  5. 香港小学中文常用字网址
  6. 汉字叔叔的汉字古代字形网站
  7. UniHan

How to cite this: If your learning and research uses the data we provide, please reference our article (we will update our publication list) or our website (Meaningfully learning Chinese characters at

Project team members: Xiaoyong Yan, Zhesi Shen, Jianzhang Bao, Richard Sears, Jinshan Wu

Send us emails