After a long absence, I’ve recently revisited James Heisig’s excellent Remembering the Kanji. Years ago, I completed the entire book, remembering the vast majority of characters rather successfully. Then, I came to Japan, lost interest in Kanji, and promptly forgot everything.
For those unfamiliar with the Heisig Method, it assigns English keywords to each Kanji. The keywords serve as a stimulus for some kind of story that you use for remembering the makeup of the Kanji. For example, even the untrained reader can tell that 明 consists of the characters 日 and 月 put together. The keywords for those three are bright, sun and moon, respectively. You don’t have to be particularly imaginative to think of a story that links those three together.
I’ve heard criticisms of this method, but it worked extremely well for me, until I gave up on it. Anyway, recently, I’ve channelled unknown sources of motivation and regained some interest in re-learning the Kanji. I found that luckily, it’s a bit like riding a bike: you never really forget. Some things stay in memory for years, despite laying there, in disuse. Others disappear, and require additional work to bring them back.
I found that Heisig does a really good job of grouping the Kanji into lessons. The Kanji in each lesson share common components, making them easy to remember. Strangely, I haven’t seen any efforts to visualize these relationships. The closest that I could find is this - it’s Kanji, but it isn’t really Heisig-related, and covers a very small subset.
So, I decided to make my own. I’ll save the technical details for later, but just briefly:
- Each Kanji is a node in a bidirected graph
- Edges from A to B indicate that A contains B (roughly-speaking)
- Groups of mutually connected Kanji form connected components
Here’s the result for Lesson 17, the one I’m on right now:
Text in red under each Kanji indicates its Heisig keyword.
If you’re interested, here’s a git repository that I’ve created for exploring such relationships. Hope you find it interesting. I know I certainly did!