How many characters is enough?

[leosmith]

JimmySeal, I've heard about 3,000 characters/10,000 words = 99% in Mandarin. Are you using the RTK method to learn the new characters?

[JimmySeal]

No, I'm just brute forcing them for now. I'm now familiar enough with kanji that most of them stick in my head without mnemonics. I might slow down and Heisig the new ones when my trip is over.

[david1978]

Sorry to say this but something smells rotten in the state of Denmark. If I understand correctly you say you are "feverishly studying" 30 new characters per day, every day? Well gosh, man! You should be a fluent Chinese speaker in less than a year at that rate! I just don't believe you know what you're talking about, sorry.

[JimmySeal]

Thank you for your repeated apologies, but they are unnecessary.

That middle part might be true, if I were able to keep that pace, but as the number of characters increases, retaining the already learned ones becomes harder, and I have slowed down to about 10 per day after reaching about 1300.

I do think it is possible to reach a high level of reading and listening proficiency within a year, and I think that with enough practice, that can be parlayed into speaking ability, maybe over a bit longer period of time.

[gato]

David, Jimmy said before that he already knows more than 2000 kanjis from learning Japanese. For the kanjis that have exact equivalents in Chinese, he just needs to learn their new Mandarin reading. So it's not so unbelievable that he can learn 30 "new" Chinese characters per day. But I agree that he can't keep this pace up because at this rate, he'll be learning 10000+ new characters in a year.

[sthubbar]

Jimmy, let's see how this challenge works out. I contend that 4000+ characters are necessary to read Chinese.

[stephanhodges]

I tracked down the reference to "anki" at http://repose.cx/anki/index.html but I don't know what "the RTK method" is.

The top Google search for "RTK Method" returns:

The RTK method is available in SewerGEMS to calibrate your sewer model and is used to generate a hydrograph based on precitation data by combining ...

I'm sure THAT's not it

[stephanhodges]

Ah, never, mind. "RTK" = "Remembering the Kanji"

[furyou_gaijin]

Sorry to say this but something smells rotten in the state of Denmark. If I understand correctly you say you are "feverishly studying" 30 new characters per day, every day? Well gosh, man! You should be a fluent Chinese speaker in less than a year at that rate! I just don't believe you know what you're talking about, sorry.

And what is wrong with becoming a fluent... well, if not a speaker then a reader is less than a year?! Some people don't have a lifetime to dedicate to this pursuit.

I'm very much like JimmySeal on this and I have a personal goal of 6,000 characters by year end. It looks like both or us have started off with 2,000+ of Japanese characters so it can be safely assumed that we do know HOW to learn characters...

I'm actually finding this quite easy and very entertaining. My biggest challenge in this seems to be memorising the correct tones.

[david1978]

And what is wrong with becoming a fluent... well, if not a speaker then a reader is less than a year?! Some people don't have a lifetime to dedicate to this pursuit.

Did I say there was anything wrong with it? I don't believe it's possible, that's all. I don't care how many Kanji you know, you're not going to be a fluent Chinese reader in a year, or two, or three...

[JimmySeal]

I don't believe it's possible, that's all. I don't care how many Kanji you know, you're not going to be a fluent Chinese reader in a year, or two, or three...

Well is there anything wrong with learning a few thousand characters to work towards that goal?

And I believe you are wrong. I think that with my language background, it's entirely possible to reach a considerable level of reading ability within a year, and be able to read practically any common writing within three years of diligent study.

And at what point did I say that I plan to keep up the pace of 30 a day indefinitely? I knew that I would have to taper it off at some point as the burden of reviewing previous characters increased, and that is what I am doing. I have nonetheless learned to pronounce about 1400 characters in 2 1/2 months, and that is a good start towards being able to read.

[sthubbar]

I'm in the mood to give unsolicited advise.

First, furyou_gaijin, way to go on your aggressive attack on Hanzi.

furyou_gaijin said:

I'm very much like JimmySeal on this and I have a personal goal of 6,000 characters by year end. It looks like both or us have started off with 2,000+ of Japanese characters so it can be safely assumed that we do know HOW to learn characters...

I wouldn't worry too much about 6000. 4000 will make you literate and 5000 will get you way above the vast majority of native speakers.

[furyou_gaijin]

I don't believe it's possible, that's all. I don't care how many Kanji you know, you're not going to be a fluent Chinese reader in a year, or two, or three...

Hm... I actually know someone who did become a speaker and a reader in a year, albeit with more spare time on his hands.

I don't suppose you'd care to explain why this goal is unattainable?

I wouldn't worry too much about 6000. 4000 will make you literate and 5000 will get you way above the vast majority of native speakers.

I'm reactive and lazy: I have a book that teaches 6,000 (and explains why you need them all) so I'm going to learn the 6,000. Will let you know how I'm getting on...

[leosmith]

There are some who don't like to pre-learn characters. It's a personal choice, and depends on what one's study plan is. But to say the method doesn't work is incorrect IMO.

Many have found learning the radicals really well and then using the "story" method very effective. I used the method to learn 20 kanji per day. I learned about 2200 kanji total using this method. Is it possible to learn 30 hanzi per day? I believe it. How about 100 kanji per day?

[sthubbar]

furyou_gaijin said:

I'm reactive and lazy: I have a book that teaches 6,000 (and explains why you need them all) so I'm going to learn the 6,000. Will let you know how I'm getting on...

What book?

[furyou_gaijin]

With all the discussions around sustainability of the method that a few of us seem to be using, it might be a good idea to give a brief overview of my own routine...

My book is called 'Cracking the Chinese Puzzle' and it is by T.K. Ann. I am using a massive 5-volume edition but there is also an abridged 1-volume version of it. The 1-volume edition is also featured in Stuart Jay Ray's video on learning languages (I recommend for everyone to look it up on the youtube, it's hugely inspirational), he literally says: 'It's a great book! After you've read it, you will have learnt 5,000 characters.' The full version actually contains around 6,000, with variants.

The approach is loosely similar to Heisig's but instead of inventing a bunch of 'lies' and assigning random names to radicals for the sake of convenience, this book actually explains the real etymology for each character, tracing it back to the Bone-Shell script, etc. It also contains examples of usage, chengyu, bisyllabic expressions, etc.

I have quickly read all the volumes once - this gave me a general understanding of how things stick together, an appreciation for the magnitude of the task and a very good idea of how phonetic components work and to which extent they are useful.

I now have a routine in which I am going through 50 new characters per day, reviewing through the SRS software (iFlash, in my case). I think even 100 per day is sustainable if one has more time to spare. Additionally, my routine involves a study of component words and chengyu. I also look up many characters in internet dictionaries and on Google and Baidu, just to see how they are used. I also use the Chinese wikipedia to get a notion of which fish, which plants, etc. are denoted with the characters that I encounter. This is quite thorough and increases the work load so I personally can't do more than 50 on a given day. I also travel extensively for my work and often have to skip days and deal with piled-up reviews. Allowing for this, I still believe I can manage the 6,000 by year-end.

I'm not making any silly stories to remember the character form as the etymology provides enough information to be able to retain it. I try to remember the readings by loosely associating them to words in other languages I know (thus, 'ji' may be remembered through 'jeans'). For tones, I'm using loose association to various moods: the first tone is something connected with machines and no feeling, the second tone is questioning, the third tone is something arcane and secret, the fourth tone is a command or something very abrupt. After playing around with various techniques, I have found that this one works for me the best, though I'm not dogmatic about it and may use another association if it readily suggests itself.

The key is also that I try to do a lot of reading and force myself to fully recognise the 'known' characters in the text.

And of course, 50 per day is sustainable because of grouping them by component. So when learning the very common 誰, I would learn 焦, 錐, 椎,騅, 雈, 鵻, 隹, 蕉, 醮, 蟭, 礁, 潐, 膲, 僬, 燋, 趭, 蘸, 噍 and 瞧. Incidentally, one will discover that all of the above only share 2-3 readings, with occasional tone variations. And etymology and the Bushou are there to provide clues to the actual meaning...

Still doesn't sound sustainable?

[Long Pan]

The government has a list of 3500 frequently used characters and a larger list of 7000 commonly used characters

Is there a website / database with these lists ?

[gato]

2500 Most Commonly Used Characters and 1000 Next Most Commonly Used Characters

http://hyacz.blog.sohu.com/32109983.html

现代汉语常用字表

7000 Generally Used Characters

http://www.moe.edu.cn/edoas/website18/info16839.htm

《现代汉语通用字表》

Some statistics:

A media survey done by the Ministry of Education and State Language Committee found that

- the 10356 most frequently used words consisted of 2463 unique characters

- 4000 words covered 80%; 11,000 words 90%; 13,000 words 99%

http://www.moe.gov.cn/edoas/website18/info19704.htm

教育部、国家语委发布2005年中国语言生活状况报告

　　5．高频词语10356条，共使用汉字2463个。

　7．覆盖率清楚显示出高频词集中的特点。覆盖率为80%时用词4000条，覆盖率为90%时用词1.1万条，覆盖率为99%时用词13万条。其余152万条词覆盖率仅为1%。

[wushijiao]

Just to share my personal experience, in my second year of studying Chinese (roughly), I became obsessed with trying to learn as many characters as possible, partly because I thought that getting to 2,500/3,000 (or some other frequently cited figure) was the key to being able to "read the newspaper or books". So, each day I would wake up with the goal of learning X number of characters. My process was remarkably similar to furyou_gaijin's (although perhaps not quite as good): I had T.K. Ann's Cracking the Chinese Puzzles (which is still a great book, in my opinion). I used flashcards. I read lots of random things....etc.

However, I eventually found that I could read the vast majority of the characters in a newspaper, and yet couldn't make any sense out of it, except for the rough gist.

What went wrong? Clearly, I had way, way over emphasized the role that individual character recognition plays in the reading process. Although the analogy is flawed, it is like memorizing a 2,500 prefixes, suffixes, and word roots and then expecting to understand a modern working language. To some degree, I think that the people who always cite numbers like 2,500 characters as the numbers of characters at which you can become fluent at, without addressing the fact the characters are just building blocks for actual words, are either English-speaking travel book writers who don't know what they are talking about, or they are Chinese people who are viewing the issue from the point of view of Chinese literacy, which is completely different from CSL literacy, because Chinese people are already fluent in the spoken language! What is much more important, in my opinion, is vocabulary, and to a lesser extent, grammatical patterns.

On the other hand, I think that the solid foundation that I built in characters did serve me well once I adjusted my learning strategy. In that sense, it was a bit like a football player in the off-season who lifts weights, does sprints, and works on overall fitness. It isn't directly related to football, but playing the game at a high level would be hard without that brute training. At the end of the day, going crazy about characters will only help you in the long-run, but it might have severe opportunity costs.

To get the best of both worlds, I think it might be good to 1) be aware of some of the deficiencies that can result from an overemphasis on characters 2) combine some of the hardcore character studying regimens (as mentioned above) with a lot of other strategies, but especially massive, massive amounts of listening. Listening is the bedrock for good speaking skills, good pronunciation, good tones, using words in their proper context, and of course, being able to understand people when they are talking. Anyway, just my two cents.

By the way, as far as how many characters does it take to read, probably not too many 2,000- 3,000 or so. That does not, however, say anything about how many words are needed to read, which a vastly more significant statistic (not that I know what that is). I do agree that 2,500 is somewhat low, but the advantage of reading is that you can always take out your dictionary and look up things up! Time can stand still for you. Time is not of the essence, unless you’re in a timed test or something. Perhaps the more interesting question is how many characters does it take to understand the radio news in real time. That requires full knowledge of the formal, written form, combined with a the ability to understanding a native speaker speaking at full speed! Also, many rare words and rare characters, that never fall into the “covers 95% of the spectrum” category, are often the most significant towards understanding the passage! Just another reason to get a lot of practice listening, I think.

[david1978]

However, I eventually found that I could read the vast majority of the characters in a newspaper, and yet couldn't make any sense out of it, except for the rough gist.

Wushijiao, I think what you've outlined above is the source of my skepticism. I personally believe in using a comprehensive approach that incorporates a heavy emphasis on usage. I only learn 4 to 5 characters per day. However, those 4 or 5 characters I can perfectly pronounce, and use in a variety of grammatical situations. I just don't see any point in vacant, brute memorization. Also, let me retract my earlier statements. I do believe you can memorize 30 or so characters per day. I was under the impression that you guys were actually LEARNING the characters. But you're just committing an image to memory. In that case, I can do 100 characters per day too.