Benefits of Memorising Individual Characters?

[Johnny20270]

I know this has been discussed to death in previous posts, but as an upper beginner/lower intermediate, here is my two cents on using character lists for SRS like ANKI. I created a new deck for the top 1000 characters as defined by Jun Da’s “Modern” frequency list. Even though, I think mindlessly going through a character list using ANKI in isolation is pointless, I do think there is a lot of merit in using a single character deck for ANKI (see prerequisites below). I see several reasons for this, namely:

1. Enforcing the non one to one association of character to pinyin.

I still keep forgetting that a character can have several tones and/or spelling and this misleads me with hanzi word recognition, even apparently well revised words, e.g.[/size]

着 [zhāo/zháo/zhe/zhù/zhuó] as in [随着 (suízhe),着急 (zháojí),衣着 (yīzhuó)]

行 [háng/xíng/xìng] as in [举行 (jǔxíng), 银行 (yínháng)]

2. Distinguishing characters which look very similar.

I was surprised that I failed to recognise differences in very common characters, e.g.

实[shí] as in [诚实 (chéngshí),实话 (shíhuà),其实 (qíshí),实际 (shíjì)] and

买[mǎi]

看 and 着etc

3. Correct tones in pronunciation (E->C).

For example, previously in ANKI, I kept failing to guess the correct pronunciation of the word以为 (yǐwéi). I was pronouncing 以 as yī. However now that I see an example of以 as in 以为 (yǐwéi), I seem to subconsciously have a mental note that the character 以 is important in this word.

4. To avoiding just glancing at a word.

I have covered about 800 hanzi words and on about 90% retention rate. However, I am pretty bad at recognising them as single characters (30% initial recognition) Often I fail on a flash card (for a word) as I jump to the answer in my head too quickly (and say “ahh yes, that’s what I meant”). I think this is because I have a tendency to either look at either

(i) the left hand character of a bisyllable word, or

(ii) a component / radical of a character, not focusing at the other part, and immediately guess the word incorrectly

Looking at individual characters makes me focus a bit more on the 2nd Character, addressing (i) above, or focus on components of characters (ii)

5. To expose myself to very common characters that you have never seen in a revised chinese word.

I was very surprised that the characters军 (102), 与(108), 战(145), 政(150), 产 (159), 斯(168), 立(197), do not appear at all in HSK4 word list (Unless I have something wrong with my matching formulae in excel!). Perhaps they are used a lot in Chinese names, no clue!

6. Increases English understanding of new words

When I look at a Chinese word. I usually go from Chinese -> pronunciation -> English meaning. Very often I know the English meaning of a word Chinese word just by pronunciation. I suspect most of us do (due to speaking listening). Seeing a new Chinese word and knowing the pronunciation gets me a long way to its meaning. E.g. just today I was presented with a new flashcard 主意. As I recognized the individual characters, I knew the meaning straight away because it comes up in Pimsleur a lot

For using a single character deck in ANKI, some prerequisites / notes are necessary, (in my opinion of course!)

• You need to know a decent amount of words (C->E) first (500-1000) so you have covered/seen several examples of words, containing that single character
  
• it should be a secondary consideration, rather that your primary one when learning words via flash cards through software like ANKI
  
• Its essential to create your own deck and it must have example words taken from your own personal known word list. This is quite easily created via a spreadsheet and a couple of lookup/match statement in Excel. Downloading a shared deck of just the top 1000 characters is not much use.
  
• There is an end in sight. I think as your Chinese improves the focus is more on obtaining a wide list of vocabulary than characters. Reading that the top 1000 / 3000 characters encompass 90%/99% respectively of all modern Chinese texts, the need for recognising and memorising individual characters becomes much less of a requirement.
  
• I foresee that the usefulness of reviewing single characters in ANKI diminishes after about 500 to 1000 characters, but the more experience people can guide us on that
  

Interested to know what the advanced people think. Agree/disagree?

Thanks

[Guest realmayo]

Agree.

On the characters versus words theme*, I think this is like study versus real-world and textbook versus tv & magazines etc etc: people will rightly warn you not to focus on either side over-much, but different people will have different preferences not just because of their own personalities but also because of their own situations. For instance a while back I could speak some Chinese from living there years earlier, and I was in the UK with a few months to kill before flying out to China to start a year's study: it made huge sense to spend those months force-feeding myself characters, for plenty of very good reasons.

*not that you look like you're trying to set up an argument about one or the other but that's where these discussions often end up going.

[Johnny20270]

... will have different preferences not just because of their own personalities but also because of their own situations

yes spot on! I was just listening to some lessons from Chinese pod, 2nd level (beginner I think). I still struggle with that as a learning source although it seems excellent for colloquial Chinese!. I far far prefer a well structured grammar book like Yip and Remmington. Chinese pod depresses me because of the non structured nature of it, but I can see how it really appeals to others. Although I force myself to do it, their teaching method (top down approach) is not natural to me at all. I am constantly (frustratingly) trying to see word order, grammar patterns and rules as defined by Yip and Remington. Just suits my personality, and probably not for the best.

[tysond]

I was finding learning words slow and annoying due to many of the issues you mentioned above, so I took a 1 year detour and learned 3000 characters using Heisig. Now that characters are a fairly easy part of learning, and I focus on adding vocabulary, learning grammar, and of course pronunciation.

Chinesepod seems unstructured, but I find if I focus on particular topics/themes, I build quite a good working vocabulary and can have discussions. It's more of a web of knowledge than a tree.

[Johnny20270]

3000 characters in a year! Well done!

Well done, I tried Tuttle, I presume itrs similar to Heisig in using stories to memoris characters . I didnlt get in well with it as I seem to spend more time memorising the stories and not sure if the same results can;t be achieved just by a combination of brute force & word examples & own stories

[tysond]

Just to be clear, I did brute force it. I just did it using the same decomposition, imaginative visual memorization, revise via SRS steps for each character, day after day. The hard part is doing it every day.

[sparrow]

I've got to disagree with this. I'm probably Intemediate/Advanced and I've never found learning characters in isolation useful.

 

A theoretical rational is that we learn better when putting things in context.

 

A more practical rational is that individual characters are not terribly useful. We all know Chinese words tend to come in pairs. The Routledge frequency dictionaries contain 5000 of the most frequently used words (and perhaps a few hundred more due to their topical lists, which contain animals, foods, materials, and so on). Those 5000 most frequently used words contain about 2000 of the most frequently used characters and, generally speaking, provide about 90% comprehension.

 

After that point, the number of new words you need to learn will decrease drastically as you sift through study materials, and will be increasingly specialized.

 

Furthermore, when you've learned five words containing the character 致 , for example, you will get a sense of its meaning and may be able to guess at the meaning of new words that use that character.

 

Compare this to learning characters: When you encounter a character you've learned in a new word, you can guess at the meaning, but you still may need to look up the word for a more precise definition—your work is still cut out for you. However, if you focus on learning actual words, it is more likely you have already learned the new word.

 

In addition, when you learn a character, you often need to attach several abstract concepts to each character. This is not only difficult, but you are forced to do this kind of learning without any context—a context that can only be provided by putting the characters into the words they form.

 

Now, if you are taking 3000 of the most frequently used characters and want to create a context for each character, then maybe you'll learn three words for every character you study—thus fleshing out how the character is actually used in Mandarin. By doing this, you are actually trying to learn 9000 words, many of which may not score high in terms of frequency.

 

 

My opinion is that Mandarin students focus on word frequency; it is more effective as well as more efficient. But that is just my opinion. We all have to find our own way.

[tysond]

The same argument about context can be made for learning words without learning sentences as context.  

 

At some point I decided to build a foundation on characters, and then sentences.  

 

From the reading I'd done it seemed to me that the concept of words is not quite so fundamental to Chinese, so this seemed like a sensible way to go.  Plus I like the Khatzumoto's writing a lot so I thought I'd give his advice a go.

[sparrow]

@Tysond:

 

You are correct. But the same thing about sentences can be said about paragraphs, and so on.

 

The way around this is to study words by frequency. Because these words appear frequently, you will encounter them in their context very often.

 

"From the reading I'd done it seemed to me that the concept of words is not quite so fundamental to Chinese."

This statement really confuses me. Could you point me to some sources you've seen about this? In Modern Chinese, words are by and large disyllabic—I don't remember the number off the top of my head, but I think it's 80% or higher.

 

Here is some proof to back this up. The HSK is a Chinese fluency exam for foreigners. You can find the vocabulary lists by scrolling almost to the bottom of this site. I downloaded the vocabulary lists using some spreadsheet equations, was calculated the percentage of Chinese words that use more than one character. Here are the results.

 

 

As you can see, as the difficulty level of the words increases, the number of multi-character words increases. Out of nearly 5000 words, 86% are muli-character.

 

The new HSK only tests 5000 words, and, if I remember correctly, the lists are somewhat generated by frequency. In the most difficult 2500 words, the number of multi-character words is 94%. You can imagine that in the next 5000 words would be over 95% multi-character. Thus, you can see that the Chinese language is very much so made up of words, rather than individual characters. (However, from what I've heard, this is not true for Ancient Chinese.)

 

In light of the above evidence, I'm quite convinced that studying Chinese by word rather than by character is the best way to go. However, I would be very interested in seeing a convincing argument for studying Chinese words character-by-character. If you have any arguments for that approach or sources that convinced you to take that approach, please let me know.

 

 

 

Sparrow

[sparrow]

If you want to verify my results, here's how:

1. Download those vocab lists. They are spreadsheets.
2. For total words in the list, in each spreadsheet, enter the formula =COUNTA(B3:B3000).
3. For total words with length greater than one character, in each spreadsheet, press CTRL+SHIFT+ENTER after typing =SUM(IF(LEN(B3:B3000)>1;1;0)), which enters this as an array formula.
4. Copy these results to a separate spreadsheet. Divide the cell from #3 by the cell from #2 to get the percentage of words with length greater than one character.
5. ???
6. Profit!

[tysond]

Thanks, I've already done this analysis some time ago for SUBTLEX-CH.  I am aware of the distribution of lengths in word lists.

I do use the word lists to help prioritize which sentences to learn.  But they are, in the end, lists of words. 

 

I chose learning characters because I found learning words was not interesting and I felt no sense of progress because I kept getting stuck on recognizing characters.  So I wanted to eliminate that stumbling block.

After that I found learning sentences to be very useful, which means the words are in context -- along with their modifiers, ordering, colocations, etc.

 

Of course I have no idea if learning some other way would be "better".  I don't care, this way worked so far, and more importantly, I didn't get bored or frustrated and stop.  

[sparrow]

Could you explain how you got stuck on recognizing characters? After all, whether you're learning words or characters, the part where you learn writing/recognition should be exactly the same—you take the vocab word or character and write it several times. The recognition should come by itself.

 

But if your way works for you, that's great!

[li3wei1]

Of course I have no idea if learning some other way would be "better".  I don't care, this way worked so far, and more importantly, I didn't get bored or frustrated and stop

 

I think all discussions of 'the best way to . . . ' will come down to this. You can drag out any number of studies on neurocognitive whatever, but in the end, learning Chinese takes time, and whatever method helps you put in that time will be the best method for you. And you may need several methods - one for when you can listen but nothing else (driving, jogging), one for when you are tired, one for when you are alert and have all your resources to hand.

[tysond]

Sparrow - sounds like you are very quick to learn characters, that's great.  Writing a few times doesn't actually allow me to reliably write and recognize the character, especially when you start getting into 1000+ characters because many of them look similar.

 

I was unable to get past about 150 characters using just "write it till it sticks" methods.  

[sparrow]

@tysond

 

Perhaps you're right.

 

I'm curious: Did you try writing the vocab maybe three times, then rewriting it later in the day three times, then again, then again before bed time? I found when I switched to this method, by giving myself time to forget and remember several times, vocabulary stuck very well and for a fraction of the time spent studying.

 

But if you tried all that and it didn't worked, then it's just a testament to how we're all different. Best of luck!