My first attempt started on 17th January 2007, just after I’d returned from a 3 month internship in Japan, and the day after I started my PhD. On this account I did 10,153 reviews which got me to kanji 1026.

My second attempt started on 21st June 2009. 1046 reviews later I was on kanji #249.

From memory, the second attempt went better because I re-read the introduction to the book and realised that the reason I had burned out under the sheer weight of pending reviews was because I was relying on word games instead of visual imagination, something Heisig warns against.

While I felt like I was making more progress, something clearly got in the way. PhD work has a way of absorbing all your free time, but this wasn’t the reason why I quit that time. Rather, it seems that while I noticed the advice against using word games, I failed to heed the advice that the images for each character needed to be vivid. Like this guy [1] I had just come up with an image quickly so that I could move on to the next.

Heisig states in his book that at 2 hours a day, 20-25 new kanji each day should be possible. That equates to at most 6 minutes per character. I would wager I spent no longer than 2 minutes on each, with most of the time spent reading the frame in the book and writing the character several times.

The plan this time therefore is to spend much longer on each character, until the image is strong, settled and rich. I will also try to incorporate other senses such as smell and touch; if this were an exercise in meditation then that is what we would do, so I will try applying it here to see if the extra sensory information can help make each image stronger and so help with remembering the kanji.

[1] Imaginative memory is the key

One kind of meditation involves visualisation. Your object of focus can become an imaginary apple, where you are at once sensing the waxiness of its skin, its weight in your hand, the intense taste of the first bite.

Those who are already familiar with the Heisig method of kanji learning may have guessed where I’m going with this. Heisig teaches us to form vivid images for each character to aid memorisation. He provides some guidelines or rules when constructing these images, but doesn’t justify them in the book to avoid distraction; we just take them at face value, trusting his judgement.

I think however that an understanding of meditation, particular meditation through visualisation may help with improving our ability to learn kanji using Heisig’s method. In meditation, we are warned against forming superficial images that we only see; we should also sense them (the waxiness of the apple, its taste). Likewise Heisig insists that we spend time with the images and make them as vivid as possible.

There is something different about the way in which I learned the kana (using Heisig’s book) and the way I learned many of the kanji. As the kanji count approached 1000 I became lazy, relying instead on making images using the words alone, with only shallow imagery. Heisig warns against this at the start of the book in fact. Perhaps much time needs to be invested in the early stages to ensure that we are visualising correctly. At my first exposure to kana I found them extremely difficult to remember, yet putting my faith in Heisig’s method I learned them rather effortlessly.

Herein lies a possible problem, and an insight into why Heisig’s methods has some detractors who appear to practically hate the method: some people cannot visualise on demand.

In visualisation there are two types of meditation, in one you conjure the image, in the other you allow images to surface from the depths of your mind; it turns out some people just get blankness when attempting to conjure up images directly.

With this insight, perhaps there is a way to practise visualisation such that our skill with it can improve; certainly welcome news if it means higher retention rates for kanji. Another interesting thought is that if you are really into your kanji study using Heisig’s method, you may in fact be meditating already!

Just for fun, try and picture something in your mind right now. Imagine an apple for example. Is it red, green? Is the skin waxy or dry? What does it taste like?

If you can’t pictures anything you may be one of the percentage of people who can’t easily visualise like that on demand. You could instead try just keeping ‘apple’ in mind and seeing what comes up.

These days I study Japanese, and when I started the first thing I did was get a book and CD to prepare for a 3 month stint working in Japan. I also got a book on learning the hiragana and katakana by James Heisig because it happened to be recommended on Amazon.

Talk about aiming low and underestimating the scale of the task! Even if I mastered the content in those two resources I’d still be stuck with virtually no knowledge of kanji (I’d be functionally illiterate) and would be confined to some pretty formulaic set conversations. Not to mention, these textbooks always hide the difficult Japanese from you so as not to scare you when you’re starting out.

The result was that I struggled constantly with learning the language and I learned an important lesson: you need input. You need to listen to hundreds, thousands of hours of spoken language. You need to read as much as possible. Drilling set phrases and words will only get you so far (perhaps about as far as I got, struggling as I did?).

After returning to the UK to start a PhD programme the first things I did was to look for a Japanese class. I quickly found a place that looked good (its sole purpose was teaching Japanese and was affiliated with the University of Manchester) and e-mailed the guy who runs the center. After arriving 30 minutes late he looked at the speeches I’d brought along (from my time in Japan, these were speeches to my colleagues, thanking them for their hospitality etc.) and told me that I couldn’t join the class as it was for beginners and so I’d be ‘disruptive’.

So, point #1 for why language classes don’t work:

1. The class will proceed at the pace of its slowest member. Somehow it doesn’t seem proper to talk about people in a class being slow, but naturally each member of the class has their own pace for whatever reason. If it was a maths class, they’d simply be stuck, need to get a tutor, and maybe do badly in the exam. In a language class however, there’s an in-built “no one left behind” mentality. If you’re sitting in a class and you’re biting your tongue to avoid answering all the questions, and you find yourself sitting waiting for others to answer, then you’re probably in the wrong class. Tip: go for the most advanced class that they’ll let you in to; you’ll thrive there.
2. You’ll spend 90% of the time listening to non-native speakers’ attempts at the language. Who’s talking more, the students or the teacher? In my experience Japanese teachers tend to conduct their classes in English. I’ve taken classes in French, German, Spanish and Russian and right from the first day the teachers were speaking in the target language at least 50% of the time. It throws you at first, and then becomes completely natural. To then do it the other way, where the only time you hear proper spoken Japanese is when the teacher corrects something or reads out a sentence, is quite painful. The end result is that you’re very rarely hearing the language you want to learn spoken as it should be.

Now, I’d say that these two points are the most damning, and any others that come to mind are either not quite as severe or quite closely related. For example take the problems that you’re given. It’s a bit like maths at school, where you’re forced to solve a set of problems from the curriculum, and told frequently that these skills will be helpful in adult life. Of course, very few people are using trigonometry or even the arithmetic of fractions in their daily life and would have been better served by being taught problem solving skills. Analagously then, in language learning we’re typically given a number of disjoint scenarios with questions about a passage maybe. We learn these examples but when faced with an actual conversation we can’t understand what they’re saying. Then when we manage to work out what they’re saying we find we haven’t had a problem quite like this before, and so end up not being able to say much.

That all sounds rather negative, and it is, but I’m yet to meet someone studying Japanese who has only been through classes and can hold a conversation or read even simple texts well. The ones who make good progress are more likely to be those who, in their spare time, watch Japanese TV, listen to podcasts, read manga or simpler books that interest them (this is the crucial part!) and have so taken measures to immerse themselves in the language. In my case I’m abandoning all classes, turning the system on its head and just reading things that I find interesting, and listening to as much spoken Japanese as possible. This is how I learned English and Icelandic, and really there’s no biological reason why this approach can’t work for adults. But, more on that later.

One point was the over-reliance on puns and word play rather than concrete mental images. This was how I learned the kana (using Heisig’s other book, Remembering the Kana) and it worked very well. Dropping down to ‘just’ clever word play doesn’t make as good use of imaginitive memory. The image needs to be clearly visualised until it is readily associable with the keyword. The other main point was: attending a Japanese class! I’d forgotten but Heisig warns against mixing this learning method with indiscriminate kanji learning. This may seem counterintuitive at first but given that the point of Heisig’s method is to divide and conquer it makes sense. By attending a class every week and doing the homework I was required to learn kanji in an ad-hoc manner rather than in the principled framework that Heisig outlines. Also doing the exercises, writing the essays and revising vocabulary all take time away from the original activity of kanji learning. Now that the A-level exams are over I am again free to focus on kanji, and this time take care to follow Heisig’s advice. Much of his advice is given without much justification and of course no citations, but as it turns out I have found evidence supporting his assertions in the literature on memory (more on this later) so I am quite happy to follow his guidelines. I’m now on the third day of studying Heisig again and have raced through 140 kanji. This is still well within the region where I was careful to make strong images and so they are all readily memorable. I’m expecting this pace to slow later and this time I’ll be careful to moderate the pace such that there is always time to form concrete images for each new character, and so finish this course finally this year!