Characters outside of familiar context: Problem, Solution, and RFC

[phyrex]

Hey guys,

I'm having the following problem:

My reading is okayish, and I can more or less recognize about 2500 characters. The problem is, that, while I can usually identify the characters when they come up in my anki cards and in familiar words, it gets troublesome when I find the characters in words that I'm not familiar with, which is starting to really inhibit my free reading. To pick an arbitrary example, I have no problems reading 收拾, but if I saw 拾 on a check, I wouldn't be able to say "Hey, I know that thing! It's the second character of shoushi!". I suspect this problem largely applies for the second character of common words, which don't appear to frequently outside of these words. The best way to handle this might be a 'structured' approach like Heisig, or skritter, or something like that. However, I've tried this many times, and I simply get too terribly bored by these things, and as a result, it won't work for me Also, writing characters is on the last place of my priority list, and right now there's a lot of areas that need the time I can save by not writing by hand.

So today I got annoyed enough about this situation, that I hacked a bit in the hanzistats plugin so it would export me a list of all the unique characters of the 5000+ sentences I've collected in Anki so far. I then imported the list into a new Anki deck, and let the PinyinToolkit plugin fill out the other fields like meaning and (most common) reading. I'm now planning to go through the resulting deck (and update it now and then with the new characters of my sentence deck), which will prompt me on single characters. This *should* make me familiar enough with the single characters without supporting context, that I should be able to recognize them in unfamiliar words. Or at least that's how I imagine it to be

Now I'd like to issue a request for comments - Does anybody see any problems with that approach, or does it sound reasonable? I just thought it'd be a good idea to ask before I'll sink real time into it

Edited March 11, 2010 at 09:27 AM by phyrex

[shivasprogeny]

While I don't think this method will hurt anything, I'm not sure how productive it is.

Modern Chinese is highly bisyllabic and only sometimes does knowing the component characters help. For example, 放心 really doesn't make sense it terms of "put" and "heart" nor does 破产 "break property" make sense. These have to be learned as "be at ease" and "bankruptcy." Under your system if you learned that 破 means break, 破除 still wouldn't really make sense because you have to learn 破除 as a whole word.

Have you tried keeping track of the words you come across that you don't understand? I think that might be more beneficial than focusing on all the individual characters.

[phyrex]

Thanks for your answer! I think you misunderstood my approach, though!

I'm always learning words and sentences first. However, say I have learnt 破产. I learn to recognize them as a unit, and all is good. But then I come across 破除, and I don't recognize that the 破 from 破除 is the same as in 破产. This sucks, because I might have been able to learn 破除 from context, if I had know both characters ( and the meaning of the word would be inferrable form the sentence). However, I didn't pay attention enough to 破 and 产 in 破产, because I can only recognize the word as a unit. So if I learn to recognize 破 without context, I should be able to identify it both in 破产 and in 破除, thus making it much easier to pick up vocabulary while reading. Does this make sense? In short: I never ever learn characters without context (I'm well aware of how useless that is ), but once I learnt a word, I need to recognize the characters it's made up from.

[valikor]

I know only half as many characters as you, but I've already noticed this problem. It's very frustrating.

I imagine learning how to write them by hand would help a lot (I know you said you're not interested in this, nor am I after a certain point), but it's actually funny because I sometimes actually fail to recognize characters that I can consistently and accurately produce by hand! (This is not so common, but it has happened before, and is incredibly frustrating. Maybe I'm just dumb.)

I don't know the answer. But, your approach does make perfect sense to me (not that I've tried it)

[Guest realmayo]

Interesting. The orthodoxy here appears to favour learning characters as part of words. But when I learn a new character -- I use Anki flashcards -- I will make one card for the character on its own, and another for the most common two-character word I can find that uses this character). I don't have the problem you describe. Of course, sometimes you can only give a fuzzy answer to Anki if shown certain characters, but I let myself judge if that's OK or not (and I always make sure I can give the correct pronunciation).

[chrix]

I kinda feel your pain. But I'm a perfectionist, I wouldn't be just content with guessing the meaning of a compound, I'd still look it up to be sure (usually I recognise all the characters in an unfamiliar word). I mostly agree with shivasprogeny on this, as they say "the sum is greater than the parts".

But I do have a character deck myself, just so I get pronunciation right, I took the latter 2000 characters from a website listing the 4000 most common characters and put them through pinyin toolkit. I also have a rare character deck I generated with a similar method like yours, but I haven't updated it in a long time. For me, words definitely take precedence over characters.

[Guest realmayo]

Another reason why I think learning characters on their own is not as arcane or abstract an activity as it might seem is the phonetic aspect: most of the time a rare character will have the same phonetic component as one or several more common characters -- with just the radical being different. So learning the less-common one does in fact keeps you more connected & tied in with the "real world" of Chinese, and indeed helps reinforce your knowledge of those more common characters too!

[valikor]

Realmayo: Correct me if I'm wrong, but it sounds like you already are doing what phyrex was suggesting doing?

And you say you don't have the problem.

So could we say that it works (at least in your case)? Or maybe it's an issue of individual thinking styles that makes it harder for people like phyrex and I

[jbradfor]

phyrex, let us know how it goes. I have the same issue, but I have come to the conclusion that at my level of Chinese, it is best to focus entirely on words, not on characters. [My level seems to be somewhat below yours, maybe 1500 characters and 2x-3x that number of words.] In the past I did focus more on characters. Recently, however, I've found it more and more difficult to distinguish characters apart if I just focus on the characters. [Recently I've been having trouble with 既 vs 即 vs 急 vs 爭 for some reason.] Focusing on words makes it much easier to tell them apart.

The other issue I see with your approach is that many individual characters have so many different meanings. So in order for your approach to be beneficial, you really need to learn every possible meaning. This seems like a difficult task.

This is not to say I don't focus at all on characters. When I'm learning a new word and I come across a new character, I will typically look up all words with the character. But I do this more to be able to learn the word with a new character, rather than learn the character itself.

[phyrex]

jbradfor, I don't "learn" anything new about the characters. I just test myself for recognition. Like "羸, yeah, I recognize that one! It's something like 'weak' from 羸弱. And it's not 赢!!". I don't try to learn all possible meanings from 羸 there (though pinyininput filled them in for me, and if I pick up another meaning now and then, all the better).

I really just want to be able to reliably *identify* the characters that I know

[jbradfor]

"identify" as in "yes I know this character from somewhere?" To what end?

[phyrex]

"identify" the character as in "oh, it's THIS character! I know it from word XX. It's usually pronounced YY and has the approximate meaning of ZZ". To the end which I've stated in my initial post.

[chrix]

Just one word of caution about pinyin toolkit: it's not so great with duoyinzi/poyinzi, as it doesn't give all the possible readings (understandable as it's not a character dictionary) but it also often doesn't give you the more common reading but whatever is first in its dictionary. If you care about this, and I do, you have to be very careful when going through your character deck...

[phyrex]

Yeah, I know, but thanks for the warning.

Since all the characters come from my sentences, where I've looked up all the unknown words, I'm usually going with that reading. I'm not going through all 2700 cards 'just to make sure' it's got the right reading, but when I'm reviewing and I'm getting a different reading than what I expect, I double check the character in my dictionary and look which sentence it came from.

[chrix]

Actually if you're interested I have a list of the most common poyinzi....

[phyrex]

Haha, you really have a knack for lists, don't you? ;) [bTW, is that chengyu thing still working?]

I'm not too interested in it, but thanks for the offer If I could stand learning from lists I would have powered through a list of the most common 3000 characters, then all the HSK words, and lastly a grammar book, right at the beginning of my studies. Unfortunately, this is nothing that works for me So I doubt that a poyinzi list will be an exception to that rule However, I'd like to know how many characters you've collected in your list! It annoys me when Japanese learners always say that in Chinese characters have in all but the most obscure cases only one reading and one meaning

[chrix]

Actually I have a dictionary with 1,338 poyinzi, but most of them are obscure readings, so that's why I started collecting the most common cases. Maybe up to 100-200 right now, but it's not at a point where I'd publish this yet, as this is still a work in progress...

The chengyu thing, oh we need to talk about the chengyu thing... will send you an email...

[phyrex]

The approach seems to do me good. I just found out that 厉 and 历 are, in fact, not the same character! ;)

[renzhe]

This is the very reason why I have always had two decks -- one with the most common characters and one with the most common words.

I learned 宿舍 and 奢侈 as words, and for the life of me, I couldn't remember which character is which, years after having learned them. It took me a long time (and the phrase 舍不得) to remember 舍.

Also, I found it easier to learn characters first, then compounds made from these characters, than the other way around.

[phyrex]

I tried it that way in the beginning, but I found that without knowing words which were using them, I never had enough motivation to learn characters on their own