Visualizing Pinyin Tones

Right, of course, and what I was trying to say is that I'm not sure if that is much more effective in the long run. I don't know if any research has been done on this?

I think I understand what you mean now, and I have similar reservations.

Memorising a character (word, etc.) is only half of the problem. Retaining it is another. I believe that, while you can memorise thousands of characters in a few months using one of the mnemonic systems, retaining and internalising all that information will likely take much longer. So, once you're done learning with mnemonics, you're not finished, you still need to flashcard a lot, or many of them will eventually fade, mnemonics or not.

Personally, I brute force vocabulary a lot and use mnemonics with the tricky ones. It's worked great for me.

I guess I can give some feedback on how it has been working for me: I do not use it very often. There are probably less than 30 instances where I applied it. I still fail those tones - but can jump right back to the story I attatched to it, and the next couple of times it usually is fine. So it is working fine - but it is not magic.

@renzhe - out of curiosity: what kind of mnemonics do you use for remembering tough hanzi?

Edited March 5, 2009 at 04:44 PM by HerrPetersen

I just make them up on the spot. The more ridiculous, the better. For example, 推 = hand pushes bird, 断 = hammer (on the right) breaking a glass pane (left). I don't know, I've forgotten most of them

Pretty nice - having a mnemonic for every hanzi might be overdoing it - but I got used to it, and don't feel it is too much extra work. Anyhow I yesterday made up the following, which sticks pretty well - hopefully it is a "learned once - never-forgotten-candidate":

嘴: when checking a *mouth* for kissability *compare* from different *angles*.

Edited March 6, 2009 at 05:46 PM by HerrPetersen

But HP, that's not 比 it's 此! Perhaps it could be changed to "kiss a mouth from this angle"?

Damn, you are right! I should probably use a bigger font - thanks for clearing it up. I had a few occasions where this kind of stuff happened to me, for instance: I was writing the bird/(Heisig - turkey) part of 唯 with only three horizontal strokes. It was enforced even more, since at that time I was still using paper-flashcards with no possibilty to check back - still I did not realize even after seeing the new Kanji with "bird" in it.

Anyhow - got a new story while using "this here" like suggested here.

Edited March 7, 2009 at 04:29 PM by HerrPetersen

I've used the Heisig book to memorize 2000+ Japanese characters. I worked out a mneumonic device that made remembering their Japanese pronunciation much easier (The ON yomi, technically based on the Chinese pronunciation); It involves chunking characters together and those chunks all "live" in a distinct location. I still remember their locations (therefore also the pronounciation) to this day (Thanks to Supermemo), and as a result I have no problems with Japanese characters, I have no problems learning new words and I enjoy reading stuff that has lots of characters in them; say what you want about the validity of mneumonics, this system has worked so well for me that I can't be convinced that it isn't valid.

My current project is to figure out how all of this work can be integrated into the Chinese language, with the added benefit of all Chinese characters generally having only one pronunciation (Thus greater benefits to reap per character, should an easy system of transfer be found). The pins all seem to be lined up, all I need to do is find the right bowling ball. I get the feeling that I'm not going to have to apply this method to each character en mass, but if a system that lends itself to quick memorization can be created, then by all means it will be applied across the board, whether the mnemonic connection seems valid or not. Forgive me, I'm kind of "thinking out loud" here, trying to work through this problem among others that are trying to work through the same thing.

I've toyed with the idea of having different tones represent different people (Fake or real), but in my own short-term experiments I had a hard recalling such details to the point that I could tell it was becoming a waste of time. Plus, it could easily conflict with the present mnemonic stories that are already set in place.

Right now I'm thinking of giving attributes to each tone. For example, the fourth tone could be a robot theme. Whatever a "forth tone" character takes up a place in the imaginary world, the actors/actresses would have robot attributes, or robots/technology could take a more prominent role as the events within the world transpire (A robot is the one that pulls the switch, moves something, blows something up, etc.). I'm looking for different ideas that are flexible enough to function as "nets," to the point that whenever you see a character you'll say "Oh yeah, that was definitely a robot character," or something like that. If it requires too much thought then it means that the system isn't good enough.

A bulk of the work is already done, I just need to find an easy way to make new additions so it can be applied to Chinese. When the results of the present experiment are apparent I'll post on here again.

Nice to hear from another Heisig graduate!

What are your goals concerning Chinese - is it just a side-project alongside Japanese or are you in full Chinese-mode now? Did you change up some Heisig-stories to fit the Chinese-characters? Can you give an example for a mnemomic for the pronounciation of a certain charachter?

Sorry for being nosey, but I am just interested.

With Japanese most of the learning obstacles are gone (Kanji, weird counting systems, grammar, etc.). While I still try to learn three new Japanese words a day (When I've made such a big time investment, and they take so little effort to learn, I might as well), my main focus is now learning Chinese.

When studying Japanese I had two "secret weapons" that greatly contributed to my fluency (And still do): the Heisig books (Along with a workable ON yomi system) and Supermemo. Supermemo (A flashcard system) is easily applicable across the "learning stuff" board, its relevance is clear. But the thing I've had a harder time applying have been the Heisig methods and stories to remembering the Chinese pronunciation. Because Chinese and Japanese use a great number of similar/same characters, knowing what many of the characters mean has been an obvious help in decoding words and understanding the nuanced differences between similar words. But the challenge of applying this in the realm of Chinese pronounciation still stands.

When learning Japanese I played around with mnemonic techniques until I found something that suited my needs (ON yomi system). Right now, Chinese seems to be a logical but challenging extension of the already acquired skills. Imagine playing Super Mario Bros. and getting really good at it; then you start to play another platform-genre game that incorporates elements from other genres into the gameplay (Perhaps an RPG element, such as Zelda II). Sorry for the geeky metaphor, but in this case you're still using the same skills as before, you just have to combine them with other stuff to succeed. After completing Heisig's kanji book, learning the pronounciation of Chinese characters has proved to be just like playing Zelda II after playing Mario; a logical but challenging extension of the already acquired skills.

But my perspective is flawed because I learned to speak Japanese first and now I'm moving onto Chinese; I've taken a road that most reading this probably didn't take. Nevertheless, from my perspective, Heisig's book for Chinese does the same as the book for Japanese did - it gave you the keys to literacy (Which I would argue can easily carry over into the spoken language). It doesn't open the door for you; opening that door takes a great deal of effort and persistence. But once the acquired skills "click" in terms of application to the language, I'm sure they will be an invaluable learning tool. It has yet to "click" with me, although I think I'm nearing that point.

Oh yeah, the results of my previous post.

So here's what I did: I took the first 50 elements in the Remembering the Hanzi book. If any of those first 50 hanzi overlapped with the kanji that I had already learned (Which was most of them), those characters were entered into Supermemo along with the pronounciation in Chinese. The flashcard consisted of the keyword being the "Question" part, the "answer" being both the character and the pronounciation in Chinese. Here's an example:

Q: Middle

A: 中 zhong1

I devised a mnemonic system for remembering the pronounciation that involved far too many mental hoops to jump through, and WAY too much irrelevant imagery. Even though I thought the imagery was fairly "basic" (in order to convey the ideas that they encapsulated), it ended up being a jumbled mess.

Although I did notice one thing: I was able to remember the pronounciation of certain characters without too much effort, and it seemed to have very little to do with the foiled mnemonic devices. Here's the only conclusion I can come to: Seeing as I've already learned the pronounciation of "character x" in Japanese (Specifically the ON yomi), learning a new way to pronounce the character is nothing more than a more abstract form of word association. Take 早, for example. In Japanese it is pronounced "Sou." In Chinese it is pronounced "zao3." Although the two aren't exactly the same, they still bear a strong resemblance.

Maybe this is another way of saying "brute forcing it," but it seems to be working. When I encounter a character that DOESN'T bear an immediate resemblance with the Japanese pronounciation, AT THAT POINT mnemonics greatly help out. But contrary to my prior expectations, what I seem to need is not a system of complicated mnemonics, but more of a "patchwork" mnemonic style; when something comes up that can't be "brute forced," a quick but logical mnemonic helps to provide the little bridge I need to connect the two ideas; if the connection is too weak, I forget it. If the connection works, then Supermemo takes care of the rest.

For remembering the tones I have four different "guest cameo" appearances; each one is distinct and quickly conveys the tone. But the point is: the system isn't being applied to every single character indiscriminately, only to characters that need a little memory boost; otherwise the mental imagery becomes too clouded with some weird visual mental 'spam.'

Hopefully that made sense to those who have read down to this point (Again, I'm kind of typing out loud, working through this next little challenge).

I'm going to keep learning more characters (Maybe I'll wait until I hit 100), evaluate the results evident in Supermemo and then report back here with my results, observations and predictions. Thanks for letting me use this space to make sense of things.

(It's all a process of trial and error until something that consistently works is found)

Nice read, littlefish! I think “brute-forcing” with occasional “guest-mnemonics” is a good way to go. It seems to be working pretty well for me, so this should make us two . (unless you you change things up again).

Here's how I do it:

Instead of assigning a mnemonic to a tone, I try to use characters that have the same pinyin and tones. For instance, today I learned 架 (jia4). I already knew the two components, so I could go ahead and make up a story. I came up with the following:

I imagine I'm a stressed-out female professional and decided to go on a prolonged in a remote location in the African jungle [1]. There, I meet Tarzan. Of course, I immediately fall in love with him and his muscular body [2]. After living with him for a while, I run out of money (life in the jungle costs more than I thought). So, I write home to my family asking them for financial as well as for some of my personal items, since I feel homesick. The next week, an airplane [3] flies over my lofty place and drops off a piano [4] and a stash of cash. I couldn't feel happier.

Notes:

[1] My code for "this is the pinyin" is the color green. Hence the jungle

[2] One could also add the meaning of "frame" by picturing his body

[3] classifier

[4] idem

A second example is 幸 (I hope you're familiar with mainland-Europe comics):

The Dalton brothers are running an illegal -cracker factory [1] somewhere in the desert. For that purpose, they kidnapped a few members of the weak . They force them make all the dirty work. Since the women's clothes were not suitable for this hard job, they had to exchange dresses with the ladies [2]. Having to walk around in female's clothes already gave Joe enough reason to be in a really bad mood. Averell eating half of one day's production made him go berserker. As always, though, Luke saves the day by busting the illegal cracker shop and sending the Daltons back to jail.

Notes:

[1] I pictured the two strokes below 土 to be some sort of underground tunnel.

[2] The Dalton's always wear green shirts. So, the ladies are now dressed in green

Here's how I do it:

架 - frame - 加、木 - jià. (for a measure words, I might also remember a sentence, such as 一架飞机).

幸 - good fortune - 土、丷、干 - xìng (although actually I remember 丷 and 干 as one component, I just can't type it out here).

Call me old fashioned, but personally I find this significantly simpler and easier to remember than stories about the daltons, green shirts, tarzan and pianos (sorry, not meaning to pick on you in particular m_k_e, but yours was the closest post on hand).

Anyway, I appreciate that different people find different techniques more/less effective than others, but I'm always amazed at why people go to such efforts to remember such long and detailed stories about things completely unrelated to the character at hand, when it doesn't need to be that complicated

In response to imron: People's brains work in different ways. For some, "just remembering" something isn't enough. I tried to "just remember" Japanese characters by writing them over and over again, for hours; but I would forget them as soon as I finished writing. I would see them in a book, and I would say to myself "I know I learned that character, I wrote it over and over, but I don't remember it." I had no connection with this piece of knowledge; but mnemonics create a connection that wouldn't be there. And when you're dealing with 2000+ characters, you logically need a more elaborate and complex system. Before trying to use the Heisig book(s), I could read but a handful of kanji. Now I can fluently read newspaper-level written material. This isn't just because of Heisig's mnemonic methods, they were only a piece in the puzzle.

Mnemonics might be elaborate and seem unnecessary, but when they work, they can be the silver bullet (And the feeling of validation you get when you find a workable solution is greatly gratifying); I can't deny their usefulness when implemented in an effective way.

Okay, on with further results:

I'm currently on frame 200 (From the Remembering the Kanji book; I've decided to finish all of the kanji treated in that book FIRST, then moving onto the characters unique to the Chinese language). Here are the current results, and I'm very happy with them:

Only one minor/significant thing in my learning methods has changed; other than that, I'm using the same combination of brute-force with light mnemonic patchwork where needed, with Supermemo providing the needed reinforcement after successful initial learning stages.

Assuming the connection between the character and its meaning is already solidified, let's take an example:

順 is for "obey." The reading is "shun4." The mnemonic for "shun" is easy enough (Someone that is SHUNNED by society), and this will likely be forgotten once the reading is cemented in the mind. But how do you know that it is the fourth tone, and not the third or second?

Here's what I've been trying to do up until recent to fix this problem: Each tone had a particular "style" or perspective to it. With the first tone, the stories played out in an RPG-style, each action being deliberate, as if chosen by a menu (Final Fantasy ATB in real life). With the fourth tone, everything took place from a first-person perspective (Think Half-Life 2). The second and third tones were even more difficult to pin down than the other two (Which is why they didn't work, I guess). When I tried to remember the reading of a character, what appeared in my head was a big jumble of thoughts (Did this one happen in first-person? Or was he targeting that thing? etc. etc.). It wasn't working, but I thought that it was the most logical and workable solution (After all, wasn't this the most significant way that one can interact differently with a world, be it fake or real?). Anyways, it didn't work. So back to more trial-and-error stuff; here is my current solution: motion.

First tone - Straight Across

Second tone - Ascending

Third tone - ?????? (This one would be most open to interpretation, I think)

Fourth tone - Descending

Whatever mental image a particular character causes to appear in my head, I simply add motion to it. If it is first tone, something goes a great distance across (Which can involve jumping, but must stay at the same "sea level." For example, it can be an object thrown across a baseball field with great force). In the second tone, something ascends to the sky, or at least somewhere really high up from where they started out (Run up a building, pop out of the ground, etc.). In the fourth tone, something descends or falls from above (From the sky, jump off of a building, etc.).

For the third tone, the current solution (That seems to be working, although I see it as being the most vulnerable variable) involves the action of taking cover, moving, and taking cover again (Think "Gears of War"-style pop and stop). If there isn't an object to throw, they usually have a water gun. Of all the different "actions" I experimented with for the third tone, this seemed to work the quickest in terms of ascertaining the intentions of the "components" of the characters. I leave this one up to you, though. No doubt better things can be used to help recall the third tone, but I find myself looking forward to third-toned characters, being very interested in implementing that situation into the character (This could just be because I enjoy "playing games" with mental images, hence an easy fallacy of my argument).

Now, back to 順. Because this is forth tone, the image simply has to contain something FALLING. For this one: the heads of SHUNNED people FALL into the river (Because they didn't obey, of course).

母 mu3 - A cow ducks in and out of cover, shooting milk with its udders.

湖 hu2 - Imagine something someone jumps out of the lake up towards the sky (the moon, perhaps). 'WHO is it,' you ask yourself.

氷 bing1 - Sliding ACROSS the ice at breakneck speed.

That's the basic system that's in place right now, and it's working wonderfully. I'll post again when I know that it's working great, not working as well, etc.

Here is what I suggest regarding learning characters.

1. If you can "just remember" it, then fine.

2. If you need help, do not write it over and over again. Instead, think about why such a character is composed in such a way, and how it relates (if it does) to the pronunciation. For example, if I had to think about 順, I would acknowledge that it is composed of 川 and 頁 (and if I don't already know 川 and 頁, I'd learn them first, as they're very common components). 順 sounds like 川 and has something to do with 頁. I can then look up the etymology, and find out that if something is obedient, agreeable, or follows a pattern, they agree with the head, which is written 頁 instead of 首. Note that I didn't say anything about a river. As for the pronunciation, I think of the sound of 順, and perhaps the sound of 川. (Relating it to another language though the "shunned" thing hinders your memory's linguistic independence, in my opinion.) Speaking has to do with muscle memory, so I might say a few words that contain 順, like 順利, 順便, 孝順, 筆順, etc. This puts the action of saying "順" into a context.

3. If that doesn't work, do whatever.

A character I had a hard time learning is 羲 (xi1, name of an emperor, breath, vapor).

1. I couldn't "just remember" it.

2. This says it's a lot like 義, and 兮 is probably the phonetic, so in order to write it, start like 義, but stop after the 9th stroke and write 兮, and then finish normally.

3. No need to make stuff up.

Total number of times written: 1. Times forgotten: 0.

One benefit of Heisig (or similar systems) is that it works from day one. You don't have to learn a skill as imron says - you do not have to know a lot of characters as Hofmann method requires.

I remember that after the second day of learning Heisig I already knew 専 which struck me as pretty complicated at the time, and after a week I could decipher "pearl river bridge" aka "pearl creek bridge" from a soy-bean sauce label. This was after 10 days and around 500 (then) Kanji with 2 hours or so study time a day.

Also one learns, in a systematic manner, the skill of decomposing characters.

So I guess imron and Hofman's arguments are valid for people who already know a lot of Hanzi - but for complete (or relative) beginners I think Heisig has a little more magic to it.

So I guess imron and Hofman's arguments are valid for people who already know a lot of Hanzi

Nope you're wrong. I started using my method when I was a beginner and was finding that writing characters out by hand was an ineffective way to learn.

It also works from day one, just more slowly than it does after lots of practice.

At the beginning you might not be able to learn as many characters in one day, but it's false economy to think like that because you're building a solid foundation of skills that make it reasonably trivial to learn new characters. Within a short space of time it's easy enough to reach a point where you can learn and absorb characters without needing to wait for the new version of some book to come out, without needing to create stories, without needing to associate a character with completely unrelated concepts/sounds, without needing to worry about different stories creating conflicts with each other etc etc. (these are all problems I have read people posting about in this and other threads).

When I read that someone can fluently read newspaper-level written material for Kanji but still needs to come up with elaborate stories to remember a single character, then to be honest I'm a little shocked, because surely by that level you shouldn't need to still be doing that.

It all comes down to what you practice. For me, my aim is to have instant recall of a character, instant recogition of the meaning and instant recognition of the pronunciation, and so those are the things I practice. Remembering all that extra information just seems to me like adding a whole bunch of unnecessary steps to the process.

Without getting into a super long-winded response (Which I will once I get more results from my experiment thing):

The whole point of mnemonics is to create connections where there are no connections.

If it weren't for the fact that "2" rhymes with "blue," 1492 and Christopher Columbus wouldn't have anything to do with one another (At least in the minds of most people). That date would be quickly forgotten, like so many other historical dates and facts.

Mnemonics are like glue that help things stick to your brain. Some things aren't naturally "sticky" for some. Take numbers, for example. Some remember numbers very easily. I don't. But some people (Even those that don't remember numbers easily) have accomplished memory feats (Remembering x digits of pi, for example) using complicated mnemonics involving towns, animals, people, and just as much bizarre imagery you can think of. But why remember all of this seemingly irrelevant information? Because it is a means to an end. It is the glue needed to stick to the brain information that is relevant to that person.

Heisig's mnemonics are simply a recipe for mental glue. Some might be able to remember characters in a seemingly more efficient way; good for them. But in order to understand Chinese or Japanese characters you MUST have a connection with them. If no connection exists, one must be created using whatever method works for you (It's not going to just "pop" into your head). Heisig's system builds a huge "alphabet" of mnemonic connections for Chinese/Japanese characters from the ground up. You might be taking the metaphorical longer and "scenic route," but if follow the map you'll arrive at the goal: an understanding of characters. This is what makes people so enthusiastic about it, I think.

And I think you are slightly missing the point. The goal of the mnemonic systems being implemented right now is not to remember the STORIES leading up to the connections, but the relevant INFORMATION that it's connected with (順 is shun4, for example). It's very interesting to find out how others create connections between two things that don't have a connection (Character and tone, reading, etc.); this has been my main barrier to Chinese fluency. If this works, it might work for others; hence why I'm posting this.

But after the mnemonic has served its purpose, it gradually fades from memory, leaving but a footprint in your mind, the relevant connection still being intact.

Language is founded upon connections. If connections simply "popped" into our heads, everybody would be speaking/reading multiple languages, but they're not. Mnemonics are a basic building block to help create the connection that wouldn't be there. And just like training wheels, they come off after they serve their purpose. I don't see what's wrong with that.

Sorry to sound antagonistic, I just get the impression that you consider people that use mnemonics in this fashion are like dumb children drawing fanciful pictures with crayons.

Gang talk: I feel ya. However I was just saying that Heisig has "magic" to it - if the excess amount of seemingly random stories outweigh the benefits in the long run is the question. In any case a lot of newbies to Chinese/Japanese who really try it will be surprised by the magic that is the speed and high retention rate. Personally I had quiet a few moments where I thought "This would have been though without the Heisig-stuff".

Some random ramblings:

I never really made a decision "now I want to learn Japanese/Chinese". During university, when I had enough time I just tried to converse a little with my Chinese floor mates (they invited me to eat, party, call local companies to see if they could sell stuff etc.). When I found the Heisig book, I first brushed of its claim to teach you Kanji/Hanzi so easily - but still tried and was shocked by the efficiency. This was the time when I decided that learning Chinese/Japanese was something on the todo-list.

ps I might change the title to "mnemonics for Pinyin and Hanzi" or sth.

Just hit frame number 300 (In Remembering the Kanji, that is), and I think I've got it down now. The whole "motion" thing is working out great. Once I hit a good stopping point I'll post more.

EDIT: I've been using a combination of those 300 characters to learn lots of themed words (Words with "車" in it, for example), and they're all sticking. Very excited, I'll know more later (I've got to go to bed).

Edited April 4, 2009 at 03:06 AM by LittleFishChan