Evil_Dragon wrote:

Ben Bullock wrote:

No, it's called bullshit.

Out of curiosity, how many Kanji do you think are necessary to, let's say, read a good book?

Just out of curiosity, I did some work counting kanji in various books I downloaded from Aozora Bunko. Here are the titles and the number of kanji:

三四郎 (no. 794): 1817
河童 (no. 69): 1226
人間失格 (no. 301): 1568
Unique kanji in above three books: 2271

Here's the computer code, if anyone needs it:

#!/home/ben/software/install/bin/perl
use warnings;
use strict;
my $file = $ARGV[0];
if (! $file) {
    die "No file name";
}
if (! -f $file) {
    die "'$file' not found";
}
my %kanji;
open my $input, "<:encoding(shift-jis)", $file or die $!;
while (my $line = <$input>) {
    my @chars = grep /\p{InCJKUnifiedIdeographs}/, (split //, $line);
    for (@chars) {
    $kanji{$_}++;
    }
}
print scalar keys %kanji, "\n";
close $input or die $!;

Last edited by Ben Bullock (2010 February 13, 4:04 am)

Ben Bullock wrote:

The only actual analysis I've done was a project where I indexed every kanji out of about ten thousand documents in Japanese. The total number of kanji used was 2,400. That's from 10,000 documents. And many of the kanji were only used in one or two of the documents, or for one person's name.

10.000 documents of what kind? To me personally that number seems pretty low, especially for such a large number of documents.

Just for fun I made a deck out of Natsume Sousekis "I am a cat", Ankis kanji statistics are as follows:

2839 total unique kanji.
Old Jouyou: 1775 of 1945 (91.3%).
New Jouyou: 141 of 191 (73.8%).
Jinmeiyou (reg): 320 of 645 (49.6%).
Jinmeiyou (var): 5 of 145 (3.4%).
598 non-jouyou kanji.

Well, one could of course argue how Natsume Souseki is notorious for his use of Kanji. This probably does not even slightly compare to any newspaper out there. However, if this is the target one is aiming for (I am), good luck.

(Off Topic, sorry)

nest0r wrote:

I know we've had some people do kanji counts in other threads, but I can't remember: has anyone done them on light novels? There's a bunch floating around *cough* raseru *cough* on Nyaatorrents and various forums...

キノの旅

1353 total unique kanji.
Old Jouyou: 1217 of 1945 (62.6%).
New Jouyou: 48 of 191 (25.1%).
Jinmeiyou (reg): 41 of 645 (6.4%).
Jinmeiyou (var): 0 of 145 (0.0%).
47 non-jouyou kanji.


ゼロの使い魔


1484 total unique kanji.
Old Jouyou: 1273 of 1945 (65.4%).
New Jouyou: 70 of 191 (36.6%).
Jinmeiyou (reg): 68 of 645 (10.5%).
Jinmeiyou (var): 1 of 145 (0.7%).
72 non-jouyou kanji.

Most publications are loaded with furigana for the rare kanji.

Newspapers aren't hard because they use lots of kanji, they are hard because they use a lot of WORDS that aren't frequently encountered in daily conversation or more casual reading material. Once you get to a high level of Japanese you'll see that kanji are no big deal, but unknown words are the problem. If everything was just a kanji problem I'd have kanken 1.5kyuu right now.

Last edited by Jarvik7 (2010 February 13, 4:39 am)

Womacks23 wrote:

Most publications are loaded with furigana for the rare kanji.

They usually do this only once (or every once in a while) per word, but certainly not every time. Maybe I have a bad memory, but I frequently forget how to read certain words by the time they appear a second (or third) time. It would be nice though if they were a little more generous with furigana.

Ben Bullock wrote:

The goal seems to have more to do with boasting about numbers than actual reading of Japanese.

Yeah, soon the day will come when I'll know the Daikanwajiten from cover to cover and look down upon thee. Wait, how did you find out about my plan?

Jarvik is right though. Why worry about Kanji or readings, vocabulary is far more important. You'll learn all the Kanji you'll ever need on the way.

Last edited by Evil_Dragon (2010 February 13, 5:03 am)

Jarvik7 wrote:

Newspapers aren't hard because they use lots of kanji, they are hard because they use a lot of WORDS that aren't frequently encountered in daily conversation or more casual reading material.

Yeah, I think this is absolutely right. Also newspaper articles are on a huge range of different topics, so the words you've picked up from one article are no help in the next article. With novels generally quite a few of the words you didn't know in chapter one are going to reappear later on, so you get natural reinforcement of them.

My personal blind spot with reading novels is character names -- I wish they'd furigana them on every use, because I can never remember them :-)

To the OP: I'd suggest that a goal like "work through the whole JLPT2 (1) vocab list" would be more useful than working through a list of readings, and I think would still match the kind of long-term measurable progress you're after.

mezbup wrote:

if you don't know how to read it it's going to break your reading flow when you look it up. Also, if a book is say 180 - 240 pages (roughly) that's a couple of kanji that are only used once PER page (on average).

You seem to have missed Womacks23's recent comment.  You said toward the beginning of this thread that one needs to know 2500-3000 kanji to read a book without a dictionary.  That's absurd.  If that were the case, Japanese would either (a) never read anything, or (b) carry around dictionaries around wherever they go, 'cause they sure as heck don't all know that many kanji.  We all know that neither (a) nor (b) is true.

Your trouble is that you think you need to look up every word that you don't know.  Show me someone who claims to know every word they see in their native language, and I will show you someone who is either lying or delusional.

Last edited by JimmySeal (2010 February 13, 6:41 am)

JimmySeal wrote:

mezbup wrote:

if you don't know how to read it it's going to break your reading flow when you look it up. Also, if a book is say 180 - 240 pages (roughly) that's a couple of kanji that are only used once PER page (on average).

You seem to have missed Womacks23's recent comment.  You said toward the beginning of this thread that one needs to know 2500-3000 kanji to read a book without a dictionary.  That's absurd.  If that were the case, Japanese would either (a) never read anything, or (b) carry around dictionaries around wherever they go, 'cause they sure as heck don't all know that many kanji.  We all know that neither (a) nor (b) is true.

Your trouble is that you think you need to look up every word that you don't know.  Show me someone who claims to know every word they see in their native language, and I will show you someone who is either lying or delusional.

I know what you mean, I don't always know every word I come across when i'm reading technical material or stuff that's very specific to a certain field and the word is jargon for that field. Other than that I can read a novel no problems and know all the words in it.

Denshi Jishos are awfully popular and if they didn't need them at all they wouldn't be all that popular. You don't see a hell of a lot of people in English speaking countries that carry a dictionary. Usually they have one at their house but don't often look in it. I've seen plenty carry them around. Give anyone a kanken test and see if they can nail the reading section 100% on levels higher than 3級... probably not? Time for the dictionary.

I think my point refers not to being able to read one book but being able to pick up book after book after book after book and have virtually no trouble. Even then you're missing my underlying point that I mention often, it's not kanji it's VOCAB.

No, you don't have to look up every word you don't know. No, you probably don't care about specialised vocab for certain fields, that's fairly normal. Yes, kanji outside the jouyou are used all the time in novels. Ergo, 2000 kanji only and you'll still need to reference things. That's if you're curious enough to look up words you don't know.

You know, this is an interesting debate. I'm going to ask as many Japanese as I can about it and see what they have to say. I do remember asking someone recently if they read a novel would there be things in there that they couldn't read and the answer was yes. So I don't think what I'm saying is totally UNTRUE which you seem to be saying it is.

Remember I'm talking about pretty much never needing a dictionary again. There's an insane difference between 95% and 99.5% and an equally insane one between 99.5% and 99.9%. I'm not talking about the first one, I'm talking somewhere inbetween the second and the third.

It's true you don't need to know every piece of information to understand the greater portion of a text but you have to admit for the stuff you skip over you don't know how to read that you aren't actually reading or understanding those parts, you're just skipping them and guessing. The brain is good at filling in blanks. I'm not saying you can't read, just saying I don't consider an incomplete understanding to be "fluent reading". Not getting a handful of words over the course of an entire novel is fine but not getting a few per page is not.

mezbup: But doing any reading at all is better than not doing anything because you're afraid that if you don't know 3000 kanji you're not going to be doing "fluent reading".  I am pretty sure that I don't know 3000 kanji but I rarely encounter characters I don't know in things I read in contemporary Japanese, either for fun or for research.  I would be surprised if I even know 2000, but I'm not sure.

Soseki's books were originally published with furigana over every kanji (as was the norm for prewar popular literature, before the Touyou Kanji list).

Also, BTW, there are around 800 kanji on the Jouyou list that do not have "official" kun readings, so there's a lot more than just 了 that only has one.  Of course once you get beyond the Jouyou list, any kanji can in theory have a kun reading -- The Kanjigen gives 4 kun-yomi for 了 -- おわ（る）, おえ（る）, さと（る）, and つい（に）.  The likelihood of seeing any of these readings outside of a 漢文 text, though, is very small.  I searched google and couldn't find anything, but I did find the horrifying 和了る, read あがる, as a Mah Jong term.

Last edited by yudantaiteki (2010 February 13, 8:00 am)

You reckon you know 1200 but passed JLPT1? I don't doubt you passed JLPT1, sounds like your reading is at a pretty good level! I think perhaps you know more than you think you do numbers wise.

If you're happy, you're happy and there's no arguing with that. IMO you only learn kanji and vocab BY reading (after you're done with "study materials").

I'd still argue a few unknown words per page isn't fluent reading to a native level. It's at the reading with a dictionary level but just skipping things. It isn't a complete understanding. If it's 1 or 2 words per chapter that's not at all bad and you could definitely call that fluent reading. If it's 1 or 2 per book then you've reached a very high level.

I skip stuff when I read for pleasure but look up what I want to know or feel I need to know for comprehension purposes. When reading news, tech articles or web pages I usually look up every word I don't know for learning purposes. I understand there's a big difference, pleasure reading should be pleasurable.

Personally I'm aiming for a very high level of reading, I don't see the point in aiming low.
Why the hell not be able to read 3000 kanji? If you read all the time and Anki whatever you come across you'll get there eventually anyway. It'll probably take me the next 11 months to get to a level I'm OK with and 24 months to get to the level I'm aiming for.