Don’t use mnemonics for everything

I have written a fair number of articles praising the usefulness of mnemonics, but now it’s time to look at some misuses and limitations. Before we look at that, however, I’d like to mention that this article is part of a series of articles related to the sensible character learning challenged (it’s still open, join if you haven’t already!):

  1. Chinese character challenge: Towards a more sensible way of learning to write Chinese
  2. You can’t learn Chinese characters by rote
  3. Remembering is a skill you can learn
  4. Sensible character learning: Progress, reminders and reflections
  5. How to create mnemonics for general or abstract character components
  6. Don’t use mnemonics for everything (this article)

Even if we only include things actually present in the modern day form, to really learn a Chinese character, we need to store large amounts of information:

  • What the character means
  • Any additional meanings
  • What the components mean
  • How the components are written
  • How they are positioned in relation to each other
  • How it’s pronounced: Initial
  • How it’s pronounced: Final
  • How it’s pronounced: Tone
  • Other pronunciations (initial, final, tone)

Yes, you could design a system to handle all the sounds and tones of Chinese and encode them in a smart way and use mnemonics to learn pronunciation of every character. Yes, you can include every brush stroke into your mnemonics (I don’t actually know how this would be done, but I’m sure it could). Yes, you can include common meanings and usage in the mnemonic, too.

Using mnemonics for everything is overkill. Image credit:
Using mnemonics for everything is overkill. Image credit:

But you’d be wasting your time. This mnemonic would need a memory palace of its own (it’s not going to be just a single picture/concept). Considering what a monster it would be, it would probably occupy the entire dungeon of that palace. In short, it’s not worth it. It’s too hard and takes too much time. It’s also completely unnecessary.

So, what’s the alternative?

The obvious solution is to use mnemonics on a need-to basis:

  • If you forget the tone, use mnemonics to remember the tone
  • If you forget the components, use mnemonics to make them stick
  • If you forget the meaning, use mnemonics to figure that out

I’ve written about how to do this earlier, so I won’t repeat that again. The important thing here is to realise this:

  • If you don’t forget the tone, don’t create a mnemonic for it
  • If you don’t forget the components, don’t use mnemonics to make them stick
  • If you don’t forget the meaning, don’t create a mnemonic to take care of that

When a mnemonic for pronunciation is overkill

For instance, when it comes to pronunciation, three are lots of clues hidden in the character and if you know where to look, you don’t really need mnemonics in many cases. This requires you to be familiar with some common phonetic components but since they are… well… common, this isn’t a problem. I’m planning a separate article about this, but for now, consider these characters: 碟,諜,喋,牒,堞,蝶,蹀,鰈. They all mean completely different things, but they are all pronounced “dié”. All of them. This is because they share the same phonetic component.

Of course, this is a convenient example, but the truth is that more than 80% of all Chinese characters are created this way. Sure, it’s not necessarily exactly the same, it might have a different tone (氧/洋, yǎng/yáng), initial (湯/傷, tāng/shāng) or final (踉/浪, liàng/làng) or any combination of these, but these are still incredibly valuable clues.

When a mnemonic for character components is overkill

Regarding character components, there are many cases where we don’t actually need to be very specific, because our knowledge of the structure of Chinese characters rules out most possible combinations. If you’re creating a mnemonic for 洋, you don’t need a mnemonic to know that water goes on the left and the sheep on the right. Three drops of water almost always goes on the left and most phonetic components go on the right. In these cases, mnemonics are just there to help you get started, the rest you can easily figure out.

For instance, my mnemonic for 昏 is based on the real etymology (sun 日 setting into the ground 氐 becomes “dusk”), but note that the modern form lacks the bottom stroke in 氐 , which turns it into 氏. I don’t need a mnemonic for this, because there’s no way I would add that extra stroke by accident. I need the mnemonic to remember the components, not every single stroke and where it should go.

To each his own

All these things are highly individual and depend on individual strengths and weaknesses, our knowledge of the structure of Chinese characters and many other factors. What I don’t need a mnemonic for, you might; what you don’t need a mnemonic for, I do. And so on. The point is that you don’t do more than is necessary. If you want to kill a chicken, don’t nuke it. It’s unnecessary and you might run out of nukes for when you really need them.

7 thoughts on “Don’t use mnemonics for everything”

  1. “If you forget the meaning, use mnemonics to figure that out” this made me think… Do most of you use spaced repetition (Anki/Skritter) to recognize (i.e. Chinese -> English) or produce (English -> Chinese), or both?

    I’ve been learning characters from the books of Heisig, where the focus in on producing the character given the keyword. English is not my mother tongue, so this creates an additional problem. For example, I didn’t know that plundering and pillaging are different things, so then creating mnemonics that take that difference into account was not so trivial. Perhaps the problem is related to learning the characters in isolation, just from Heisig’s book.

    This brings another idea forward: Learning the characters and words after you’ve seen them in context. In practice this would mean reading texts and adding the unknown words into Anki/Skritter, and creating mnemonics for the difficult ones. This is a tedious process to do by hand. It’s possible to make it easier by creating a vocabulary from a given text. It can be done using a word segmenter (ICTCLAS or FudanNLP for example) and a bit of computer programming.

    For example, I can take a story or movie subtitles from somewhere, segment it into words, and get a list of all of the words in the text arranged according to their frequency in such a format that it’s easy to add to Skritter. Now the crux of the problem is finding suitable texts. I like reading comics, but this really needs the words to be in text format, not pictures.

    I’ve been reading this site on and off for a long time now, and I think that it would benefit greatly from a list or blog entry about different reading/listening/watching resources that many people could have access to. Maybe people could describe to you what they have used? Good resources are of course a matter of opinion.

    1. Hi Eemeli,

      I will definitely try to provide more resources in the future. I have done that for music, but I should do something similar for TV, books, online reading, videos and so on. The problem is that writing one article like that easily takes ten times longer than writing any other article and I just don’t have that much time. I’ll see what I can do.

      About it being tedious to record words, I think that’s necessary and not necessarily bad. I have manually added most of the 22000 words/characters in my Anki deck and I don’t regret doing it. I think adding the word, looking up sentences and so on is valuable in itself. It’s much more than just mechanically adding something to a list.

      (I corrected the link in your comment and deleted your own revision of it, hope this is okay.)



    2. Even for native english speakers, the difference between plundering and pillaging is pretty small :-) Maybe a real pirate would know but not your average speaker.

      For this kind of thing, first of all I don’t over-do my reaction to remembering the wrong one – you are 95% correct and that’s pretty awesome.

      Second, I build a mnemonic using wordplay – e.g. “pillaging a drugstore for pills” or “plunder downunder” so as to uniquely identify which variant of the same basic meaning I meant.

      For “By means of” I’ve had to include a bi-sexual in the story because I just couldn’t remember the “By” to start with, and there are so many abstract expressions to get confused with.

      Over time you get better at distinguishing and the right variant seems to pop into your brain without too much trouble.

      I agree with you that context is super useful – frequently once I have seen a character in the wild, it’s much easier for me to remember. Context is your friend.

  2. Thanks for the heads up. My biggest problem with learning Chinese at the moment is that I can’t memorise lessons just by reading through them. Currently I’m reading Assimil Chinese With Ease which provides cool little dialogues with translations and explanations, but when I try to sit down and read through them I get bored because I can’t remember anything the moment I shut my eyes. Mnemonics such as in Heisig’s book have been awesome for me since it makes me visualise a scene rather than gaze at the paper. The infinitely best way for me to learn something is to explain it to someone else, so perhaps I could imagine explaining each sentence in my head to someone…

    Anyway, have you seen this: ?
    Could you please provide any comments on if you think it would be a good idea. It seems impressive to me, but requires an investment.

    1. Yes, I have read that article and for me personally, it’s a very good example of using nukes to kill the chicken. Most students don’t need that and would be much, much better of learning more about phonetic components and simply studying characters more in general. You could use something like that for tricky cases, but doing that for all characters is exactly what my own article tells you not to do!

      1. I agree with Olle. For me, many of the characters’ pronunciations are easy to remember without any extra effort. That’s because I see them often enough in context. The tricky cases often involve remembering the wrong tone, in which case you could just add a simple element to the story (e.g. color, or I like to add feelings: calm: tone 1, wondering/confused: tone 2, nuisance/annoyance: 3rd, anger: 4th). For the rest, you can just develop something else. Those characters will account for less than 10% of all of them.

        A good way to see which tones you know and which not is to read some easy texts aloud. I often notice that the text I could understand 100% was hard to read aloud, meaning that I should focus a bit more of the pronunciation.

  3. I tried Heisig’s book but it was so much work to remember the mnemonic that I quickly dropped it. Now I just try to remember some aspect of the character that’s a trigger and will hopefully help me remember the rest. But the more characters I learn, the harder that gets. Spaced rote repetition, besides actual real-world use, is the only way I’ve found to learn them.
    (Learning to write characters is especially hard because no one hand-writes much anymore, and Chinese keyboards use pinyin to find the characters, so there isn’t a lot of opportunity to actually write Chinese characters in real-world use. I’ve talked to college educated Chinese who say they can’t remember how to write a lot of characters anymore because these days they just type them. It’s like how my teen daughter’s English spelling is terrible because she’s become totally reliant on spellcheck and never writes stuff by hand anymore anyway.)

Comments are closed.