Chinese Character Mnemonics

John “I build an entirely new weblog every two months” Biesnecker has just put up an interesting article on his newest new weblog, My Chinese Life. The article deals with mnemonic devices for memorizing Chinese characters. (You probably want to read it before you continue if you want to understand fully what I discuss below.)

John talks about how he remembers the characters (“manure”) and (something like “business”). For the former, he uses the actual meanings of the character’s constituent parts: 米 and 共. For the latter, he assigns his own meaning to the character in order to remember it. Both work.

This mnemonic device thing is something I’ve been thinking about for a long time… I think most Chinese would be quick to suggest that students learn the actual etymologies of the characters in order to remember them, but in many cases, this task is ridiculously complex and just places more of a burden on the student. It reminds me of my Calc 2 teacher’s response when we asked if we could use a formula sheet for our tests: “why would you need a formula sheet? If you forget a formula, you can just derive it on the spot.”

An example of the complexities involved in using etymologies to memorize characters is the way that a human “hand” is written as part of many characters. Can you identify the seven components which each mean “hand” in the following four characters?

友 祭 授 手

If you can do it without some really lucky guessing, it means you know your Chinese character etymology. The problem is that the forces which created the modern Chinese character set were often not systematic, or at least not systematic enough to make memorization by etymology a simple matter.

The logical solution, then, is to use a mnemonic system not totally based on etymology. There are two approaches to this. One could take the “semi-etymological” approach to mnemonics by using the real Chinese meanings of the character components in mnemonic devices. For some characters (such as John’s example of 粪), this is not hard to do at all. Your mnemonic may very well be very similar to the logic of the character etymology. In this case, etymology is your ally. For other characters, however, this proves quite ineffective.

When studying the etymology and analyzing the meanings of the component parts doesn’t work for you, what can you do? Well, you could use some kind of rote memorization method, or you could try the other approach: the “to-hell-with-etymology” approach. In this approach, you make up your own meaning for the component parts (like John did for 商).

I first read about this method while I was studying in Japan. I discovered it in an excellent book by James W. Heisig called Remembering the Kanji. Heisig’s method for associating meaning with form readily abandons the original meaning of characters’ component parts if the original meaning does not aid in systematic memorization through simple mnemonics. It works, although the specific system Heisig developed for Japanese limps in a few areas.

I think that this concept of devising a self-consistent mnemonic system for remembering Chinese characters is the holy grail of Chinese character pedagogy, considered impossible (for very good reasons) by many. It’s a problem I hope to tackle down the road. There’s just got to be some systematic method of learning a large quantity of Chinese characters that’s better than rote memorization.


John Pasden

John is a Shanghai-based linguist and entrepreneur, founder of AllSet Learning.


  1. Damn, I only got 6! I didn’t know that 爫 was also a form of hand (though it sort of makes sense). My teacher at 吉大 spent a good part of a week going through radicals and common character components talking about their origins. I guess it finally paid off. 🙂

  2. Here’s what I did:

    1) Learn the first 100-200 characters by rote memorization. There is really no way around this. I strongly recommend John DeFrancis’ “Chinese Reader” textbooks. They systematically repeat characters in later lessons, so the review is “built in”. You can just read them cover-to-cover with very little need for backtracking or staring at pages for hours. Also, the order in which new characters are presented follows their frequency of occurence pretty closely.

    (I got through the whole six-book series, which I think teaches 1200 characters by the end, in about two years, reading a little at a time as I rode my exercise bike :).

    2) Learn all the character components and how to analyze complex characters. There are plenty of good books on this. I would strongly emphasize that most characters can be divided into a semantic radical plus a phonetic. (Most of my native Chinese teachers did a good job of describing radicals, but they never taught the very useful phonetics!).

    Once I got over the initial “hump” of 200 or so common characters, I switched from pure rote memorization to a radical/phonetic approach. (well it’s still rote, but it’s a lot easier to remember 授 = 立手旁+受 than just a bunch of strokes).

    3) At about 1500 characters you get to a second “hump.” Learning gets difficult because you already know all the common ones, so the ones you don’t know are going to be rare. Cementing them in long-term memory is hard since you probably won’t see a given new character again for a long time. I think you just have to force yourself to write down new characters and review them constantly. Don’t just look them up once and move on.

    (this is where I am now… I think I’m around 1800 or so. I can ALMOST read most newspaper articles, although it’s really frustrating because I’ll know all the characters except for a few nouns here and there, which always turn out to be the most important ones… It’s no help if I can only read “the received a strongly positive report on the recent operations of .” 🙂

    One more thing… I have to make a special effort to learn characters that people use in first and last names, since many of them don’t show up very often in regular text.

  3. BTW Tokyo Damage Report has a hilarious take on convoluted character etymologies here:

  4. Dan,

    Yeah, the way I learned my characters was basically the same process you describe. It will get you recognizing all the basic characters, but I don’t think it’s especially helpful in training you to write them, or in helping to push your vocabulary into the “educated Chinese person” range.

    As for the link, that was AWESOME. Thank you! I heartily endorse that link. Seriously.

  5. I don’t think any system of rule-based mnemonics will get anybody’s vocabulary into the “educated (whatever country) person” range. That takes massive comprehensible input. As for writing, the vocabulary has to be used in context. Spending an hour a day writing isolated characters won’t do much if you never write them in daily life situations. The problem is that almost no CSL learners actually have occasion to write in Chinese on a daily basis. Unless you’re a full time student, or are already at a very advanced level, it would be pretty tough.

  6. the opposite is also true for Chinese people using mnemonics to learn English, which is also one reason for their horrific pronounciation of some things. not for written English, but for spoken English. when i first heard things like 三块肉为你妈吃 (thank you very much) or 美丽克雷斯马斯 (merry christmas – i think those are the right characters) i thought they were a joke. but they’re not. those and many more are used everyday by Chinese people learning English.

  7. The Tokyo Damage Report article was indeed very entertaining, but unfortunately knowing Chinese takes some of the humour out of it. For example, many of the seemingly ridiculous radicals are there for phonetic, not semantic reasons, but then Japanese words have been assigned to them and the original phonetics are lost. I’m suddenly very glad I’m learning Chinese and not Japanese.

  8. Todd,

    Could you share with us the source(s) (books?) that translate the radicals in Chinese characters into phonetic sounds? Many thanks in advance.

  9. Frank, it’s widely known that a very large proportion of Chinese characters can be separated into two parts, one part related to meaning and one part related to sound. Unfortunately, two words which were (presumably) pronounced the same way a thousand years ago may have shifted pronunciation since, so this does not actually mean you can always guess the pronunciation of a character by breaking it down in this way. Any intermediate Chinese learner could give you dozens of examples — simply phonetic connections which they have noticed by themselves, not read in a book.

    For example, the Tokyo Damage Report article mentions 皮, which in Mandarin is pronounced pi2 and means skin or leather. The character 疲 is pronounced exactly the same way and means fatigued — it is composed of 皮 underneath a radical meaning illness. 皮 also appears in 坡, 婆, and 破, which have nothing to do with skin or leather but are pronounced po1, po2, and po4 respectively. It also appears in 波, 玻, and 菠, which are all pronouced bo1 (note that the only difference between the sounds “p” and “b” is that the former is aspirated).

    The article tells us that 波 is pronounced “nami” in Japanese and means Poland. It is presumably influenced by Chinese, in which Poland is written 波兰 and pronounced “bolan”, an approximation of the country’s name in English (c.f. Polska, the country’s own name for itself).

    So, there is no consistent way (at least in modern Mandarin) to convert radicals to phonetic sounds, but noticing these phonetic connections (where two characters differ only slightly in pronunciation) can at least be of some limited help in learning Chinese. Furthermore, it explains why combining 波 with the female radical forms a character meaning “old woman” — the 波 character is there for phonetic, not semantic reasons. In Japanese, however, these patterns are all lost.

  10. Frank – see John DeFrancis’ “The Chinese Language: Fact and Fantasy.”

    It’s an academic book but for a language learner it provides a very helpful counterpoint to the (often misguided) things native Chinese say about their writing system.

  11. I think studying Rick Harbaugh’s 中文字谱 (Chinese Characters–A Genealogy and Dictionary), which is online at, is a great way to grasp the phonetics of the Chinese writing system, because characters with the same phonetic component are all grouped together in Harbaugh’s dictionary. The online dictionary is good, but I recommend the book.

    Nothing beats being able to leaf through the book. You can observe helpful rules of thumb like this:

    In the 4000 characters that make up this dictionary, all the characters with the 主 phonetic component are pronounced zhu4 except for 主 itself, which is of course pronounced zhu3. Outside the first 4000, I have only come across one more exception to this rule: 拄 zhu3, which means to lean on a stick. Does anyone know other exceptions?

    I think one successful strategy is to learn the rules, memorize the exceptions. But inducing rules takes effort. Well-spent in my opinion.

    Meanwhile I keep asking myself how a career academician (he’s now an Economics professor at Princeton, I believe) found enough spare time in his studies to compile this dictionary. He writes in the foreward that the dictionary “began as a distraction from my master’s studies in economics at National Taiwan University.” There may well be other models of laowai greatness out there besides Da Shan!

    Zhongwen zipu is based on 说闻解字 (Shuowen Jiezi), a dictionary written around 100 A.D. The Shuowen etymologies are apparently not all accurate, but they represent an important source in Chinese scholarly discourse.

    I need mnemonics for tones, which I often forget. For a long time I couldn’t remember that 局 is pronounced ju2 and 居 is pronounced ju1. I kept mixing them up. Then I came up with a mnemonic: To work your way up in a bureaucracy, you can sleep with your boss. So 局 is ju2, the rising tone. (Don’t ask, it works for me.) Whereas, home is the place you sleep, lying down horizontally, so 居 is ju1, the level tone.

    I have a lot of other mnemonics. Many of them are rated R, I’m hesitant to admit. But they say the stranger or more interesting your mnemonic is, the more you’ll be likely to remember.

    • The dirtier, the better to remember, according to Rocketmemory and several others.

      Do you have a book for ready-made mnemonics? I am shopping around for one.

  12. Todd,

    Thanks for your explanation. I always thought that the radicals included only the messy lines and dots in the first half of a character.

    Dan Maas,

    I’ll look for the book. Hopefully they have it the Foreign Language Bookstore. But then I’d rather live in the sino-fantasy land where I could learn up to 50 Chinese characcters every day and would never forget any of them…


    I’ve been visiting since John mentioned it in his Mnemonics blog. I’m seriously deprived of sleep.

    ju2 is my language means a male’s genital and I hardly make this particular mistake. Still can’t figure out your mnemonic for this one though.

  13. Frank,

    As I said, don’t ask 🙂 Up, you know, up in the bureaucracy. We all want to move up up up… oh, whatever.

    Yours is much better. What language is that? I’m very curious.


  14. Laksa,

    Oh…. got it now. But then at least 500 characters, in my opinion, would also quality for this mnemonic.

    The language is Thai. Lots of common words in Mandarin have an obscene meaning in my language (since Thai’s also a monosyllabic language). For example, imagine saying “ass-hole” (xie xie) to Chinese people when expressing your appreciation on a daily basis.

  15. How I learned enough characters to read novels and newspaper articles in Chinese.

    I sat down with a newspaper/book and a dictionary and read them. That’s it. I had alraedy had about 1 1/2 years training in Chinese, so a lot of charcters had been drilled into my head through homework exercises and the like. Then I just started reading… no fancy classes, no special teachers/tutors (sometimes I’d ask a Chinese person to explain something if I couldn’t figure it out at all), nothing but reading material and a dictionary. It was slow, but (assuming your brain works the same way mine does) you will teach yourself all of the reading skills you need.

    When you do extensive reading, you tend to come across the same characters over and over again. The trick is don’t get too upset if you can’t remember a word or character. If you don’t remember it the first time, you will the 10th, or the 20th time you come across it. As for stuff that you only see once ever, who cares? Also, I think its good to us a paper dictionary and not some kind of setup involving automatic computer lookup (yes, I know paper dicts. are slow, etc., but you have to pay your dues) because looking things up is one way to get familiar with the structure and pronounciation of the characters and other characters that are similar to it.

    In short: don’t wait until you think you know enough characters to read a newspaper, just start reading it, dangit!

    A small caveat: I can read Chinese pretty well, but I probably can only write about 1/4 or less of the characters that I can recognize. It’s really frustrating… it seems two very different parts of the brain are used for these two skills, and they don’t talk to each other very often. Since a lot of my time studying was done without a formal classroom setting, I didn’t practice writing much. Still, there is nothing bad about knowing how to read Chinese!

  16. Frank, that’s funny. I didn’t read your comment until just now.

    My mnemonics for tones just match some image that has a directional component with the meaning of the character. The more outlandish the image, the better.

    David-I learned (am learning) how to read the same way. But studying for the HSK, I realized that I really needed to know how to write as well. So I’m doing daily writing now. I find there is strong synergy between writing and reading, now that I’m forcing myself to do it. One of my biggest challenges would be tones. I can read an article, but I can’t necessarily read it out loud very well. Through daily writing practice, I’m finding the tones are getting easier and easier. It’s just repetition drilling in a very thorough, low-leve way with characters in context.

    I’m reminded of my music teacher’s advice–practice to your weaknesses and try to practice more than one thing at once.

  17. Hello John. This may be setting a record for tardiest comment ever posted, but as it appears you are still accepting input regarding Chinese Character Mnemonics, I hope you won’t mind if I pitch in my three cents’ worth.

    You note:

    “(T)his concept of devising a self-consistent mnemonic system for remembering Chinese characters is the holy grail of Chinese character pedagogy, considered impossible (for very good reasons) by many.”

    Yes, such a task must be consigned to the realm of the impossible. I don’t know which of the “very good reasons” for this you had in mind, but to state the case, I would note for starters:

    1) Components of a compound character sometimes serve as abbreviated forms of more complex components.

    2) Components are sometimes employed as shape indicators rather than for their normal semantic functions.

    3) The semantic function of a component may be either its independent meaning (ex: 甘 signifying sweet/good in 香 and 甜) or its conceptual sense (ex: 甘 signifying contain/conceal in 某 and 酣).

    4) Graphic changes sometimes mask original forms.

    5) Some principal contemporary meanings are borrowings from other terms.

    Nonetheless, I believe that from a pedagogic standpoint it may be worth attempting to elucidate the rules and exceptions. The semantic values of the radicals are consistent enough. The key therefore is whether or not we can, as we did with 甘 above, narrow down the semantic values of a given element to no more than two: the independent sense and, if different, the conceptual sense. If so, we will have taken a giant step toward imposing the degree of order necessary to make a mnemonic approach pedagogically viable.

    To elaborate just a bit, let us take as an example the 主 phonetic component discussed by laska, originally a pictograph of a stolid lampstand. Obviously, if we try to attach the signifier “lampstand” to this component at every appearance the result is absurd. But what happens if we take the lampstand in its sense of a massive, stationary object (as the character’s word family connections suggest we do), and apply the signifier (or mnemonic) “stationary” to this component? Our mnemonics become:

    住: Stationary + person (= dwelling)
    柱: Stationary + tree/wood (= pillar)
    注: Stationary + water (= water poured into a container, rendering it stationary)
    駅: Stationary + horse (= horse tied at a staging post)

    In this way, the student associates 主 either with “stationary” or with “master,” as s/he does “contain/conceal” or “good/sweet” with 甘. Naturally, it may be that the principal contemporary meanings for any given character are precisely those indicated by the radical + mnemonic, or they may be extended/associated senses. In either case, we have a consistent mnemonic on which to hang these meanings, which in itself is no small achievement. Applying this principle to other components will take us a long way toward an effective mnemonic approach to Chinese characters, at least with respect to the traditional forms.

    For all that, and to sum up, the nature of the Chinese writing system precludes the development of a mnemonic system valid for more than about 3/4 of the characters. That leaves the question: Is it worth creating a mnemonic system when we know it cannot apply 1/4 of the time? I’m inclined to say yes, but then again I’m terribly biased. I wonder if you or your readers would care to weigh in on this question.


  18. i dont think theres any getting around learning a lot of radicals by rote. for various definitions of a character, the best i find is to learn ancient basic meaning and trying to figure out how the various meaning came to be. get to know the general sound in various words with a character and it usually will hint at the sound. nothing will help with tone except learn it in a multi-character word the rythme. however, sometimes characters, the most common word has a tone change (sandhi). ive had chinese people argue the tone was the tone-changed tone, so the most common may not be the best to learn the tone. ive also found the phonetic part has some meaning – i assume phonetic parts werent randonly choosen.

  19. SeekTruthFromFacts Says: October 28, 2012 at 8:06 am

    Since this post is still accepting comments almost 7 years later… it’s worth noting that Heisig (with Richardson) has now published ‘Remembering the Simplified Hanzi’. My friends have used it and recommend it.

Leave a Reply