Tower of Babelfish
The Blog

In search of more efficient ways to learn languages.

  • Recent Posts

  • February 2014
    M T W T F S S
    « Dec    
  • Archives

  • Meta

  • Tags

  • On dealing with non-phonetic languages like Chinese (And how to deal with homophones in any language)

    Posted on by

    I keep addressing this question in various forum/comment discussions, but I should discuss it here, because it addresses a lot of issues in a lot of languages.  The question is:

    How do I deal with a language that uses a pictographic alphabet like Chinese?

    I’ve had a fascinating experience seeing how all of this works in Russian, my first really “foreign” language. The main adjustment I’ve needed to make is to be extremely careful with my card design - in Italian, I could put a picture of a dog and “il cane” on the other side without problems. In Russian, that doesn’t work; I absolutely need two cards - one with the dog picture on the front and another with the word собака on the front. If I don’t have both cards, I can’t remember the words longer than a few weeks before I forget them again. If I did an Asian language (particularly Chinese or Kanji in Japanese), I’d probably need at least 3 cards per word (probably 4). Interestingly, this corresponds exactly to the Foreign Services Institute’s (FSI) estimates for language learning - Russian is supposed to take twice as long as French, and most Asian languages 4x as long.

    For those studying other languages, Chinese has a few special difficulties:

    • The alphabet is a series of pictographic symbols that correspond to specific meanings of words (one symbol for “10″ (十), another symbol for “stone” (石)).  These symbols don’t tell you anything about the pronunciation.
    • Every word also, naturally, has a sound, which can be represented by a special learning alphabet used by foreigners called “Pinyin”.  So our stone, 石, is shí in Pinyin.
    • Chinese is a tonal language, so shí, shī, shǐ and shì are 4 totally unrelated ‘words’, even though they share the exact same vowel and consonant.
    • Chinese has a great many homonyms, like “cymbal” and “symbol”.  Stone is “shí”, but so is the number 10 (and a whole bunch of other words).  You can see this played out to its extreme in the poem “The Lion Eating Poet in the Stone Den” (recording of it here):
    In English:
    In a stone den was a poet called Shi, who was a lion addict, and had resolved to eat ten lions.
    He often went to the market to look for lions.
    At ten o’clock, ten lions had just arrived at the market.
    At that time, Shi had just arrived at the market. …
    In Pinyin:
    Shíshì shīshì Shī Shì, shì shī, shì shí shí shī.
    Shì shíshí shì shì shì shī.
    Shí shí, shì shí shī shì shì.
    Shì shí, shì Shī Shì shì shì. …


    Difficult, no?  First off, you must must must be able to distinguish the 4 tones without difficulty or else you’re screwed.  You can get away with forgetting to learn about pronunciation for Spanish or French (it will slow down your learning, but you can still do it).  This is just impossible in Chinese.

    Secondly, Chinese (and other similar languages), poses some interesting challenges in terms of card design because of the homophones.  I would make (at least) the following cards for ”10″ (十, shí) and “stone” (石, shí):

    Card 1: meaning 1 - sound Front of Card: 10 Back of Card:  shí

    Card 2: meaning 1 - symbol Front of Card: 10 Back of Card:  十

    Card 3: symbol 1 - sound Front of Card: 十 Back of Card:  shí

    Card 4: meaning 2 - symbol Front of Card: picture of stone Back of Card:  shí

    Card 5: meaning 2 - symbol Front of Card: picture of stone Back of Card:  石

    Card 6: symbol 2 - sound Front of Card: 石 Back of Card:  shí

    If I still couldn’t remember it (again, I suspect that Chinese might require 4 cards per word for an English speaker, given the FSI’s estimates on Chinese difficulty compared with, say, French), I might have another card or two about the symbol itself (whether the symbol is the base of other symbols or is based upon other symbols, etc. - Chinese characters have component parts, and breaking those parts down and memorizing them in Anki can be helpful).

    You’ll notice that there’s no sound - meaning or sound - symbol card.  In any language, if you get a bunch of definitions for a single word, it becomes tedious to try and recite the whole list.  The word sign in English can be a noun, in which case it can refer to these sorts of signs:

    or signs of Aliens:

    or more abstractly, as in ‘signs of fatigue‘.  And then there’s to sign, which usually refers to making a signature, but could also refer to speaking in sign language in the right context.  Making one card in Anki with “sign” on the front and 5 definitions on the back is an option, but if you mark the card as wrong every time you fail to recite a list of 5 definitions every time you see it, you’re going to be seeing (and hating) that card every 4-7 days for years.  You need to design cards to make sure that you get them right most of the time.

    For our sign example, make 5 cards with pictures/definitions on the front and sign on the back.   If you want to train recognizing the word sign, then you can give yourself added context:

    Front of card: to sign (“He signed the bottom of the application”) Back of card: to produce a signature

    Front of card: to sign (“He signed ‘Hello’ to his friend”) Back of card: to communicate in sign language

    Our 6 Chinese card examples have enough redundancy to train all of the needed aspects of 10 and stone.  Each language needs its own structure to capture all of these aspects properly, while training only one of them (say the connection between 石 and stone, or 石 and the sound shí) at a time.

    10 thoughts on “On dealing with non-phonetic languages like Chinese (And how to deal with homophones in any language)

    1. Mark


      I think that a lot of the time consuming part of learning characters can also be resolved by studying the radicals of characters. If you learn all 214 then you can “create” your own meaning mnemonics to help you remember the cards. In that way the character itself becomes a “picture” of sorts to help you remember the meaning of the word.

      So, for myself, phase one would be to learn the radicals. for radicals the pronunciation is not as vital since no one actually talks about them specifically (well, usually, but it isn’t as important when first starting) so you could do cards with a picture on one side and the radical on the other to remember the meanings. that gives you the meaning, but it doesn’t actually tell you how to say it. For example the radical 口 (‘kou3′) means “mouth”. So you would focus on remember that the character/radical has the meaning of that picture of a mouth.

      At the same time I would set up a deck of cards to develop pinyin pronunciation. So, on one side of the card is the pinyin fundamental phonetic elements and the other side of the card is an audio of the pronunciation. They are just sounds so no need to memorize any meanings. There are a few hundred of these, so do-able in a few weeks. These would go pretty quickly as it is a focus on pronunciation and phonetics, not on meaning or vocabulary.

      After you have those two pieces in place, then learning characters and vocabulary could be pretty quick. You could do it with just 2 card types:

      1-2) pronunciation (pinyin) picture
      3-4) pronunciation (pinyin) character

      The first set will emphasize the immediate call-and-response of seeing the picture and fluently knowing how to say the word.

      The second set would allow you to associate the fluently known word sound with the character, which you can remember because you know the radicals that comprise it’s meaning.

      Actually, both sets are essentially a combination of “sound + picture” cards. The only difference is that the picture of the chinese character is somewhat coded based on a pre-determined set of radical images. But, like any code, once you have it worked out you can create whole complex image correlations in your mind.

      At least, this is how I would approach it if I was going to do Chinese again from scratch…

      Thanks for the post!

      1. Mark

        Sorry, it didn’t take my symbols. It should be

        1-2) pronunciation (pinyin) + picture
        3-4) pronunciation (pinyin) + character

      2. Glenn

        That’s also really helpful Mark, but I was wondering if that would apply for Simplified Chinese as well?

        I’m still new to Chinese writing, so I’m still figuring out what my approach should be. I’m kind of split between learning Traditional / Simplified.

        Traditional has an appeal because like you said, a character is made up of radicals(was it?)and it would see to be easier to remember.

        Simplified, has more of a practical appeal, as it seems to be the most widely used. It seems logical to invest time into it first, but I’m thinking it would take longer to learn since, the radicals don’t have meaning by themselves, rather, they are a short form of the traditional…


        1. Willow Young

          I am a Chinese girl from the Mainland, and I grow up learning the simplified Chinese. As it is simplified, it does loses something of culture vein. But we still benefits from it when it comes to efficiency and convenience. I never learn the traditional Chinese, because it is much more complicated and I can recognize it. It takes you greater effort and time to learn the traditional than the simplified.

    2. gwyner Post author

      Some neat resources that will land in the Chinese page on the next site-wide update:

      This site has a simply wonderful app for teaching the tones:

      For quick pinyin-Character entry, use

      For more user friendly Pinyin-Character conversion, use

    3. Andrew

      Very interesting Gabe, I’ve always struggled with the kanji and tried different methods. However Japanese isnt a tonal language and that many cards would be over kill.
      Perhaps a card with a photo on the front of a rock and th back with the character, then another card with th character on the front with the hiragana on the back.

      1. gwyner Post author

        Hi Andrew! It’s not so much about the tones as it is about the Kanji part, and building a connection between the Kanji symbol, the meaning of it, and the sound of it. You might be able to get away with 3 cards - one Kanji-Kana card, for example.

    4. Pingback: 8 Ways to Create Better Flashcards

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    Note: For anti-spam reasons, if you put a link in your post, it will be moderated. Post *another* comment (without any links) referring to your other comment and I will look through the moderation queue and save it from certain doom