So, I recently took the time (about 2½ months) to work my way through volume 1 of Heisig’s Remembering the Kanji. The following is my review, which I also plan to post on Amazon.
Bottom Line: It works
Okay, so here’s the bottom line. It works. It does precisely what it sets out to do. I’ve been studying the Japanese language since I was 13 years old; it’s now been about 20 years of studying the language. It hasn’t been entirely consistent; it has often been a few months on (full-bore), a few months off (after burnout). Before I worked through RTK, I was probably familiar with around 300-400 of the most frequently-used kanji. I could never seem to get much past that hump, it always felt very much like an uphill battle, and even though I’d hit the foot of the hill running full-bore, I’d never quite make it to the top before, exhausted, I’d start to slip downhill again.
I’d beat my head against a Japanese reader for a while, which, though organized in such a way as to introduce a few characters at a time, instead of leaving you to deal with whatever characters you may happen to find in other reading materials, still progressed fairly rapidly, and was also fairly outdated. Then I’d beat my head against some “real” reading material for a while. Then I’d try to ramp myself up on gradeschool-level texts, so there’d be fewer kanji to deal with; but this doesn’t necessarily help so much, since kanji can often be a key to understanding a compound word’s roots, and help a good deal in learning the words from which they’re built.
The problem is, every time I came to an unfamiliar kanji, it would put a hard stop to the flow of my reading. I don’t know the meaning, I don’t know the pronunciation, I probably don’t know the next couple of characters after it, and I don’t know the word in which it’s appearing. I can’t continue reading until I’ve spent a while studying each individual character, how they’re pronounced in this context, and what the word means in which they appear.
Since finishing volume one of RTK (about a week ago), reading is dramatically easier for me. I still don’t know how to pronounce many of the characters, but since I’m already familiar with the character (which at this point has become like an old friend), it’s easier for me to attach a pronunciation to it, since I’m no longer having to learn the pronunciation, and the writing, and the meaning (maybe—sometimes it’s necessary to attach new meanings to old friends). Progress moves much more quickly; and often, even if I haven’t learned how to pronounce a string of kanji, I can tell right away, from the Heisig keywords, and from surrounding context, what the word’s meaning is. The first time I encountered 怒っている in some text after Heisig, I knew instantly what word it was (おこっている, roughly “is angry”), because I was already familiar with the Japanese word (but not its writing), and the Heisig keyword associated with that character was “anger”. Similarly, words like 簡単 (かんたん, simplicity + simple/not complex = easy) and 閉じる (とじる, to close) are immediately clear at first encounter, without having to look them up (though doing so to learn the pronunciation is advisable).
Even words whose meanings don’t happen to match the particular keywords I learned, such as 設定 (せってい, establishment + fix (in place) = “preferences/settings” (for computer programs)?), or 削除 (さくじょ, plane + exclude = delete/remove?) are easy to remember, and deepen my understanding of the characters’ true meanings (削 = plane, but also “to whittle”). Even learning 弾 as “bullet”, and then later discovering its use as “play (an instrument)”, isn’t a problem: I already know how to write it, and one of its meanings, so it’s easy to add the new meaning. Easier than learning it without context and without familiarity with its primitives (弓/bow + 単/simple), and trying to learn it amidst a sea of other graphically unrelated characters surrounding it. Basically, just having something that takes away about one and a half (how to write + an approximation of the meaning) of the three or four things I usually have to study at once when learning a kanji – how to write it, what the character means, how to pronounce it in this specific context, and the meaning of the whole word or compound in which it appears – eases the process for me tremendously.
I wish I’d found and studied this book years ago, as I’d be much much further along in my understanding of Japanese at this point if I had.
What it does not do
But let’s be clear: completing this book (or even this series) is not the end of your journey—not by a long shot. Having completed this book, you can’t claim (or at least shouldn’t, though many do) to “know” ~2,000 kanji characters. You do not. You do know “how to write” about 2,000 kanji characters, and you know a meaning for each one (not necessarily the only meaning, or even the most usual meaning). You don’t know how to pronounce a single one of these newly-learned kanji in a single context, unless you already knew beforehand. (If you proceed to volume 2, you will find a system to organize your learning of the “on’yomi” of the various characters, primarily based on signal primatives that give a strong clue to characters’ pronunciation—I’ve heard many people say you don’t need volume 2, and this is true, but I personally find it useful enough to have. Just take what you can use from it, and ditch the rest; don’t feel obligated to do it exactly as laid out.)
So you’ve learned the kanji 生 as “life”. Good for you. But how about its meanings of “student”, or “fresh”, or “birth”, or “breathe”, or “grow”? Is it pronounced セイ or ショウ, or maybe it’s なま or う.まれる or い.きる or は.える? Well, if you haven’t learned what it means and how to say it in each of a variety of contexts (all commonplace), you can hardly claim to “know” it, can you? (If you’re panicking at seeing how confusing it can be to know what a single character means and how to pronounce it, please relax: 生 is a bit of an extreme example; while there are several characters that have a wide variety of meanings or pronunciations, most have only a couple, and in fact, many “kun’yomi” are shared across multiple kanji—which one should be used depends on the nuance intended, or context.)
Completing RTK vol. 1 is not the end or even the middle of your kanji-learning experience (unless of course, like me, you were already middle-ish in your kanji-learning journey). It is the beginning: it serves as an excellent foundation (but only the foundation) for proceeding to learn real meanings and pronunciations in a variety of contexts, which is something you can only really get by reading plenty of material.
Note: The majority of complaints I’ve seen about the Heisig system seem to be that it doesn’t do various things it’s not trying to do in the first place—possibly because some of the people who complete the system claim that it does… “Now I know 2,000 kanji characters!” …no, you don’t. The other common complaint I hear is that no one who finishes this book goes on to gain an intermediate-to-advanced understanding of Japanese. This is silly, as in order for this to be true, the book would actually have to have some property that prevents you from further study. Pssht. In any case, nearly every time this challenge is issued, someone steps forward as a counter-example.
Do not kid yourself that all you have to do is read and visualize each of the kanji in order, and you’ll have learned to recognize 2,000 characters at the end of the book, without ever spending any time to review the ones you learned before. A few reviewers have complained that by the time they got to the end of the book, they’d forgotten all the previous characters from the rest of the book. Well, I mean, duh? Heisig not only never said you wouldn’t have to do any reviews, rather he outlines a specific system for you to use in reviewing them. So, um… review ’em.
Actually, I recommend ignoring his system for creating (and using) flashcards, and using a Spaced Repetition System (SRS) such as Anki, or http://kanji.koohii.com/ instead. Anki has several sets of flashcards for RTK that you can download right from within the application, and kanji.koohii.com is geared specifically for RTK. You will not only need to review, you will need to review a lot. You will become frustrated at how quickly you can forget kanji, or at least pieces of kanji, and how certain kanji (fortunately just a handful for me) keep slipping from your memory over and over (tip: if the “story” you’re using isn’t working, use a new one—however, some keywords may be inherently difficult to make associations for). The point isn’t that RTK eliminates the need for repetitive review (he does state that repeated writing of kanji is unnecessary for learning, but that’s different; either way, though, take that advice with a grain (or more) of salt), but that it significantly reduces that need (you need to review, but instead of reviewing entire character forms, you’re mostly reviewing stories, plus infrequent idiosyncratic changes to primitive forms, or unusual primitive positions).
Things it does poorly
Alright, now for some complaints about the book. Honestly, RTK sucks (it just happens to suck way less than any other method I’ve tried). The keywords chosen for kanji are frequently very poor choices (IMO). I imagine I’ll never have a clue as to why the keyword “junior” was chosen for 徒. In several cases, the English keywords themselves are obscure, and I have to look them up in a dictionary. “Decameron”?
Likewise, not enough effort is made to ensure the student chooses a helpful connotation of the keyword (Okay, the keyword is “mould”. Is that “mould” as in “there’s mould on my bread”, or is it “mould” as in “to shape” (and what’s with that British spelling for “mold” anyway 😉 )? Is it “gain” as in “he’s gaining on me”, or as in accquiring something?) Effort is made, but usually for keywords I didn’t actually need much help with, and for far too few of thsoe I really did need. Using the keyword in an example sentence for each character (or something) would have been appreciated.
Also, the “stories” used in the book are frequently very, very poor for visual association. Often they are obscure and rambling monologues with only light connections to the elements in the kanji. I very frequently replaced them with my own visualized connections, and was quite happy when Heisig finally stopped providing his own clumsy narratives.
Wait, didn’t I basically criticize the implementation of every one of the key features of this system? Why yes, yes, I did. Why do I still recommend this book, then (you may ask)? Because in the end it doesn’t matter so much that he didn’t do any of these things well; in the end, that he did them is really all that matters. He provided unique keywords to associate with each primitive and each character, so that you can easily review from keyword-to-kanji. If you don’t like ’em, you can swap ’em with your own (just be sure to check that they are unique – use the index at the back of the book to see if your keyword is already taken). He teaches the characters only after all their component pieces have been learned, which vastly improves the learning experience—instead of learning character forms, you’re learning character components, and just putting them together. As for the stories, you can always substitute your own (and, for the majority of the book, are required to in any case).
Of course, substituting your own keywords would require you to already be familiar enough to realize that the one chosen for you isn’t helpful. But the truth is, even if you learn a keyword that has nothing to do with the most common meaning (or even really any meaning) you’re going to see this character used for, you have the character in your arsenal. Once you’ve learned 徒 as “junior”, even if “junior” has little to do with anything, you still recognize the character, are comfortable and familiar with it. If you should then learn to think of it as “disciple” (キリスト教徒 = Christian, イスラム教徒 = Muslim, 聖徒 = disciple/adherent), it’s not remotely hard to replace the keyword you learned with a new meaning (which is what you’ll have to do anyway for many characters, even when the keyword you’ve learned is an actual meaning of the character).
Basically: the primitive keywords suck, and the stories suck, but you don’t need the stories, and the keywords are only a temporary association anyway that you’ll later either expand on or replace outright. The book does a good job of teaching you how to write the characters properly, and illustrates the differences between printed and written forms; and most of all, it presents everything in an order that streamlines learning.
Shortcomings to the system
The system itself has a few disadvantages which are worth mentioning, even though in my opinion they are crushingly outweighed by the advantages of this system.
First, a significant portion of your energy in reviewing and associating the characters with keywords, is that many of the keywords are confusingly similar. This is an unavoidable consequence of trying to map each character to a unique and individual keyword, since many kanji have very close meanings (which are often used to reinforce eachother when they are paired to make a kanji compound word). “Exam”, “examination”, and “test” are separate keywords. “Admonish”, “criticize”, “rebuke”, and “censure”. “Shoulder” versus “shouldered”. “Marriage”, “matrimony”, and “marry into”. Some keywords differ only very, very slightly. During review (per Heisig, always keyword-to-kanji, never the reverse), I’ll sometimes get mixed up and write the character for a similar but different keyword. I’ll then have to devote some time into focusing on what connotations the different words have that I can add to my stories to better distinguish them. This is energy I would not have to spend in learning kanji “normally”; I would learn the characters just from the contexts in which I see them used, and be able to distinguish them just on that basis, even when they have essentially the same meaning. I wouldn’t have to concentrate to think which character was associated to some specific meaning; instead, I’d just know which of the similar characters I “meant”, and use that.
Additionally, this system comes with a condition: it only really works if you work through it all at once, to the exclusion of other study methods. I don’t think you can effectively combine this with other systems simultaneously, and Heisig himself says this at least once in the introduction. This means that you have to work through all 2,000+ characters before you can begin making any use of any of them. This can be a daunting task to contemplate, and this is only for the worse since as far as I know, this system will not be effective if it is interrupted. If you drop it partway through, you can’t pick it up again where you left off, unless of course you’ve been dilligently reviewing those keywords that you had learned up to that point the whole time you were away.
For my part, I had beat my head against other methods for quite long enough to be motivated to work all the way through without stopping, and I’ve been rewarded with an excellent foundation for continuing my kanji studies.
Comparison to Kanji ABC
It’s worth pointing out a similar system for learning the Kanji: Kanji ABC. It takes a very similar approach to learning the kanji, and in particular focuses on the same key concept behind RTK: learn the primitives first, and build your knowledge of the characters from that. In my opinion, it also has a tendency to choose more useful keywords for the primitives than RTK does. However, it suffers from two shortcomings that really prevent it from being as effective as RTK, in my opinion: first, it does not teach unique, reviewable keywords for the characters themselves; only for the primitives, so you really can’t use it in isolation; you’d have to study each character thoroughly (using external means) in order to really retain any information about them. Second, it only demonstrates the printed forms of kanji, which can differ significantly from the written forms, and doesn’t really provide great coverage in general on how to write the characters.
I would love to see someone completely rework this system, and perhaps choose better keywords, and address some of the other problems I mentioned above. However, it still remains at this time, the most effective system for quickly gaining a solid repertoire of characters, and at the end of it, you really can read Japanese much more effectively. You obviously can’t read without effort and further study, but the difference in ability is well worth the 2 to 4 months you will have spent in study with RTK.