Reflections on Remembering The Kanji

So, at the time of writing I am at the end of chapter 45 (Kanji 1756) of "Remembering The Kanji" by James Heisig, and the end is in sight. This isn't the first time I've worked through it, unfortunately, but I'm quietly confident it will be the last. So, I just wanted to list a few things I've figured out doing this book that might be useful to other people going through it.

First off, I should probably explain why I think I failed the first two times I worked through it. The second time is easiest to explain: I rushed it. More specifically, I rushed through it all in one month in the summer of 2010, which means I did roughly 2042/30 ≈ 68 kanji a day. In his introduction, James Heisig cautions you not to set out too quickly[1], and it should be obvious why: If you don't take your time, you will be, by necessity, using poor quality images and stories for the kanji you come across. This bites you later on when you are trying to use simpler kanji as primitives in more complicated ones.

My advice is, if you can do 15 a day, that's probably enough. Yeah it will take you four or five months, but it's a nice workable pace even if you are kind of busy. If you have the time to do more, great, you'll get done a little earlier, but if not, you are still going to get done a hell of a lot faster that many other methods for learning kanji.

The first time I failed, in 2009 or so, was not due to rushing. I did take my time, but I'm not sure I could have kept it up even if I hadn't eventually used exams as an excuse to not keep up my reps. In hindsight, my stories were frequently lazy, abstract, and worst of all, sometimes taken from copied from other people on the web.

Some people will take issue with that last one, indeed there are sites like Reviewing the Kanji that allow people to share stories and some people will have success with that. Fair play to them, but, having them available, I used them as a crutch, and not trusting myself, abstinence from them was the way to go. I don't believe I've been harmed significantly by this. If you do want to use them, my advice is to be picky. Make sure you are picking them because they really are memorable, not because you are lazy.

So that's why I think I failed, and why I think I am not going to fall down in the same way this time, but it's not my only advice. I'm not really sure how to structure it, but here goes anyway.

I really suggest getting into a good SRS habit. This almost goes without saying, but it's for precisely that reason that I'm saying it. Pick an SRS. I like Anki, but any of the popular ones will do you. If you can pick one with a smartphone app, like Anki does, so much the better, since you can do it on the move. And try to do all your reviews every day. If you can't do reasonably do them all in one day, no sweat, just do as much as you can, even one, since it is more important to keep the habit up, than it is to finish. Don't be afraid of a "backlog"; it'll go down in time if you keep at it, and killing yourself just to finish your reps is only going to make you want to do them less.

At the beginning of chapter 27, Heisig suggests that you use the image of a specific person for 人[2]. The idea being that since the notion of "person" pops up a lot in stories, it is very easy to get mixed up and start adding that primitive where it doesn't belong. This is equally true for many other common primitives like water. While I wasn't perfect with this, but when struggling with a story, I frequently found it handy to fall back to a "stock" meaning for those sorts of primitives, e.g. in kanji like 治, the three drops of water were leaking out of a hose.

Finally, Some people have criticised Heisig[3] for his choice of keywords, and it's easy to see their point. Some of the words he uses like brocade aren't used frequently in "normal" English, and it can be hard to come up with sufficient differences for keywords like cape (埼), promontory (崎), and headland (岬)[4].I have struggled with some of the kanji for various kinds of tree.

However, I'm not convinced it is all bad. Just because you are learning Japanese, doesn't mean you couldn't stand to use a bit of practice on various English words, and somewhat strangely, some of these stories were easier for me than common, but abstract, words. In particular, take brocade and godown. A brocade is a kind of patterned fabric, and it is easy to see in 錦 a white towel patterned with golden threads with no effort. A Godown is a kind of dockside warehouse. Armed with that knowledge, I found the kanji 倉 almost pictographic: It is a dockside warehouse, the mouth allowing ships to enter and exist, and the gate houses a crane for loading goods onto the ship. Don't fear the keyword :)

I'm not sure how useful all this will be to other people, but hey, it was good to get it out, and to get this resolution experiment underway. See you next week.


[1] While he also suggests that studying full time it should be possible to do in four to six weeks, I'm not convinced that it would be possible for me to do it in that time frame.

[2] For reasons I don't quite understand, I used Igor, the hunch-backed assistant to Dracula, but it seems to have worked out fine.

[3] Intellectual honesty dictates I should provide some links, but I've been doing this for way longer that I intended already, and it's hard to get some "representative" links. Just take my word for it :)

[4] I'm still not entirely convinced I remember my story for promontory as well as I do the pain of trying to figure one out.