I'm currently working my way through RTK1 and I'm starting to wonder whether I 'm gonna do RTK3 or not.

My ambition with studying Japanese is to sometime in the future live in Japan and function in that society as a Japanese speaker. What are you guys experience with the added "extra" knowledge you get from RTK3, is it more of a personal interest of knowing as many kanji as possible or is it actually necessary to master Japanese properly?

The kanji in RTK3, for the most part, are quite rare. Once you get beyond a certain level of kanji, studying additional kanji isn't going to improve your Japanese that much, so it's basically up to whether or not you like studying RTK.

Just study the extra kanji that you actually encounter in words you learn. The RtK 3 kanji are not frequent enough to be worth studying beforehand.

I'm going with the consensus here and i'd also like to add that by the time you've finished RTK1 you're well equipped to learn any new kanji you come across.

I learned the kanji for ガン (癌　= cancer) and I was able to learn it no problems and can write it fine.

It's definitely more of a personal choice thing I think. It's nice knowing 1000 more but invest that time in learning to read or speak will do you wonders.

I have the same problem, I'm at 1400 kanjis and wonder if I'll continue or try the sentences. I remember most kanjis (more than 90% remembered everyday) and I think I'll continue with RTK3.

This thread has a few lists of kanjis to learn after RTK3 (kanjis for names, JLPT1, ...) I personally believe that having a good kanji base is essential (Who am I to say this? I'm just a beginner ) but whether you should set this base at 2000 or 3000 kanjis is your choice.

Good luck anyway.

Edit: my "progress spreadsheet" tells me that I would complete RTK3 in 46 days, I think it's a very short time if you count all the bonuses: 1945 essential kanjis + the new ones, names, JLPT1, ...

I say come back to it later. Once you're at a high enough level, systematically learning "only" 1000 kanji that you've likely seen a number of already seems simple. Once I'm done with JLPT1 next month I'm thinking of finally going through it. Then again there is kanken and jbiz to look forward to...

Jarvik7 Wrote:Once you're at a high enough level (...)

I'm confused because you may be right after all. I don't know what gagge_tah would do but here's something that could be done:

- Read RTK1 (all the jōyō kanjis but one)
- Add this kanji from RTK3 (which is a jōyō kanji too): 璽
- Add this list of "new kanjis" for 2010

I found those infos in this thread. Do you think it's OK?

mypapa12 Wrote:

Jarvik7 Wrote:Once you're at a high enough level (...)

I'm confused because you may be right after all. I don't know what gagge_tah would do but here's something that could be done:

- Read RTK1 (all the jōyō kanjis but one)
- Add this kanji from RTK3 (which is a jōyō kanji too): 璽

I have never seen that kanji used in an actual Japanese text. I guess there's no point not learning it, but you don't have to go out of your way to add it either.

Honestly I think that once you're anywhere above about 1200 or 1300, you're in a great position to just start learning the kanji that appear in the reading you are interested in. Obviously that's not compatible with RTK unless you do RTK lite, but 2000+ should be plenty. You'll never learn 100% of kanji that you will see for your whole life, so at some point you have to just start learning the kanji you need from the readings you're interested in.

In another thread there was a frequency count done on the entire Japanese Wikipedia, that's a great thing to do because Wikipedia covers so many different types of information and genres. The results were pretty much the same as for newspapers; 1300 kanji or so covered 95% and 2000 covered 99%. Going to 3000 was like 99.4% or something along those lines.

Seems like I'm just gonna leave it at RTK1 + additional I may stumble upon. Very nice to know that both newspapers and wiki doesn't use more than ~2000. Now I just gotta decide what Anki vocab list to stick to. Currently using a JLPT1-4 "complete" list with ~8000 words. But it's seems to be missing some of the "official" JLPT vocab... such a pain in *ss to crossrefence and update different lists all the time.

I was planning on going through RTK3 once I finished RTK1, but started sentence studying along with my other structured (textbook) studies. Soon I found myself finding unknown kanji out in the wild that I would have to learn, most of which weren't rare, but not common. If you are wanting to study kanji only, then you might want to tackle the new 常用漢字 as outlined here: http://en.wikipedia.org/wiki/J%C5%8Dy%C5%8D_kanji

I have a list of kanji that I am slowly crossing off (and adding to my anki deck) as I find them in the wild - these are Kanji from JLPT1 that are not 常用漢字. RTK1/3 numbers are listed as well.

此 2043
髭 2047
俺 2058
尤 2087
稽 2088
或 2091
僅 2116
凄 2133
剃 2146
匂 2147
嘩 2160
噂 2161
咳 2162
喧 2163
喉 2164
叩 2166
嘘 2167
呪 2169
吠 2170
吊 2171
噛 2172
噌 2177
叱 2179
垢 2187
嬉 2207
尻 2220
崖 2227
挨 2248
揃 2255
拶 2261
撫 2270
撒 2271
惚 2277
憧 2280
湧 2284
沙 2296
汰 2316
溜 2324
濡 2329
釜 2332
狙 2344
芯 2368
蘇 2380
蓋 2388
蒔 2395
這 2414
隙 2426
忽 2431
腎 2453
股 2454
膝 2461
肘 2464
腫 2467
枕 2472
椅 2480
眉 2583
碗 2591
稀 2611
裾 2624
箸 2640
筈 2646
紐 2661
繍 2665
蛋 2683
謎 2709
誰 2712
貼 2736
貰 2737
賑 2739
蹴 2742
醤 2746
鍋 2757
鍵 2761
闇 2777
霞 2779
鞄 2785
頓 2788
頃 2791
頬 2792
餌 2797
馳 2805
爽 2855
殆 2871
覗 2875
丼 2884
匙 2891
於 2900
壷 2904
那 2960
炒 (fry)
呟 (mutter)
睨 (glare)
籠 (basket)
茹 (poach)
其 (that)

If you have built a solid momentum from Heisig and you feel that you could maintain it for another 1000 kanji, then I think it's an entirely personal choice on your part. I agree with brianobush's point about perhaps tackling them in a more precise/selective method rather than ploughing the fields of Heisig once more?

Here's the wikipedia thread:
http://forum.koohii.com/showthread.php?tid=3216

These are the statistics quoted:
173 kanji make up 50% all kanji in Wikipedia.
454 kanji cover 75% of all kanji in Wikipedia.
874 kanji cover 90%
1214 kanji cover 95%
2061 kanji cover 99%
2456 kanji cover 99.5%
3489 kanji cover 99.9%

Now of course this doesn't mean that the 2061 there are the same as RTK's, but the general rule holds true that the more kanji you study, the less improvement you're going to see from studying additional kanji. Especially if your grammar and vocab is still at a beginner or intermediate state, doing reading to increase those is going to provide a much greater benefit to you than taking the time to study 1000 additional kanji that make less than a 1% difference in the coverage of kanji you will see.

Now, that's just talking about plain benefit, without taking your interest into account.

yudantaiteki Wrote:Here's the wikipedia thread:
http://forum.koohii.com/showthread.php?tid=3216

These are the statistics quoted:
173 kanji make up 50% all kanji in Wikipedia.
454 kanji cover 75% of all kanji in Wikipedia.
874 kanji cover 90%
1214 kanji cover 95%
2061 kanji cover 99%
2456 kanji cover 99.5%
3489 kanji cover 99.9%

Now of course this doesn't mean that the 2061 there are the same as RTK's, but the general rule holds true that the more kanji you study, the less improvement you're going to see from studying additional kanji. Especially if your grammar and vocab is still at a beginner or intermediate state, doing reading to increase those is going to provide a much greater benefit to you than taking the time to study 1000 additional kanji that make less than a 1% difference in the coverage of kanji you will see.

Now, that's just talking about plain benefit, without taking your interest into account.

It sucks that 100% sits somewhere at 50,000.

50,000 is a pretty gross overestimate; I know that's what Morohashi has in the Daikanwa Jiten but that's a dictionary of classical Chinese (with Japanese definitions), not a Japanese dictionary. Maybe if you're talking about all writing ever written by Japanese people, including classical Chinese, you might be able to reach something like that if you count all the 異体字 as different characters. (Note that the vast majority of the characters in the Daikanwa Jiten, possibly as high as 70% or higher, are simply obsolete variant forms of other characters.)

Wikipedia would only have at max somewhere in the 6000's because that's all the JIS-1 and 2 set can display.

EDIT:
From the thread it looks like 6241 characters total: "Of the 6,355 kanji in KANJIDIC (JIS X 0208), all but 114 were in the frequency list."

EDIT 2:
Note also that frequency statistics for rarer characters, especially in a case like this, can be kind of misleading without actually looking at how the characters are used. Today's featured JP wikipedia article has 77 non-Jouyou kanji in it. But that's at least in part because it's about "Asian Dust" so there are a number of Chinese place names and some personal names (given readings in the article), and some discussion of older sources with readings given for the old kanji. So in that case you don't need to literally know all 77 non-Jouyou kanji in order to read the article.

I was talking about the extremity of every kanji ever used at any point in time including all it's variants and now obsolete forms.

I think 6000 would cover 99.99%

Hmm, this talk of diminishing returns makes me wonder if there should be a further "refinement" of RTK. What makes this reasonable (and doable) is the existence of the material and the ability to sort it.

**Kanji**
(includes additional kanji for primitives)
RTK Basic - First 555 Kanji of 2001KO or the JLPT 4 and 3 Kanji.
RTK Lite - All 1110 Kanji of 2001KO or JLPT 4, 3 and 2
RTK - adapt for new Jouyou added.
RTK3 - adapt to cover all Kanken 1.5 and lower.
Kanken 1

**Grammar**
Basics - Basic Tae Kim, Ch. 1 and 2 of UBJG, Others?
Low - Tae Kim Essential, Ch. 1-4 of UBJG
Inter Low - Tae Kim Special and Advanced, All UBJG
Inter High - Kanzen Master 2
Advanced - Kanzen Master 1

**Vocabulary**
Basics - Core 1 and 2
Intermediate Low - Core 3 through 12 sorted via KO2001, upto 555th kanji.
Intermediate High - Core 3 - 12 sorted upto 1110th Kanji of KO2001
Area of Interest - Learn as required
Advanced - upto 2001st Kanji or Entire Core series, KIC, Tanuki

So, if you find yourself getting higher on one of the areas, maybe it's time to switch gears to another area.

PS: My rankings are very off the cuff estimates. Do not take serious on what I consider advanced.

Let the numbers of that Wikipedia list speak for themselves:
There are incredible 8.916 kanji listed from Wikipedia
RTK I covers 98,4%
RTK III covers an additional 1,2%

If you want to "optimize" your studies after RTK I according to that Wikipedia frequency list you can get 1,4% with the same amount of additional kanji (965). Means RTK III frequency is not that bad – it is just that the air is really thin in that region.

Or if you are more efficiency driven you could e.g. look only for the most frequent 100 kanji not covered by RTK I and get already 0,53%. Most of them are covered in RTK III.


Numbers apart, learning kanji beyond RTK I without having a useful Japanese word connected to it, does not make a lot of sense. Looking at the word examples in RTK III, I would say that this is probably the weakest point of RTK III.

Last but not least, don't worry: Whatever route you take, important none RTK I kanji as 頃 or 誰 you will learn anyway.

Nosing through brianobush's list, there's a dozen or two that stick out to to me that I've learned the readings of from various anime or Niconico vids or elsewhere online. 此れ and 其れ I learned watching 化物語 (as well as 噛む, heh), of course 俺 and 誰 are really common, 凄い and 嬉しい I see written in kanji every now and again, and a few others.

I don't think I shall ever forget the meaning of 崖, though, because of this video (shortly after 7:25):

http://www.nicovideo.jp/watch/sm4198408

腹筋崩壊注意ｗ