Nayr's Core5000 deck (Frequency Dictionary of Japanese)

Nayr182 Wrote:Here's the list of how many kanji you need to know in order to have covered which percentage of kanji-use (on average) in about 1300 novels:

up to 10%: 11 kanji
up to 20%: 20 more kanji: 31
up to 30%: 33 more kanji: 64
up to 40%: 50 more kanji: 114
up to 50%: 74 more kanji: 188
up to 60%: 103 more kanji: 291
up to 70%: 148 more kanji: 439
up to 80%: 222 more kanji: 661
up to 90%: 348 more kanji: 1045

GREAT quote Nayr182, but you left out the most interesting numbers! Here they are:

"up to 95%: 377 more kanji: 1422
"up to 99%: 768 more kanji: 2190"

According to this talk on language mastery by a "polynot" computer scientist
(which I think is a treasure trove of information and totally worth watching at 1.25x speed), you need 98% comprehension for "pleasant free reading". That's somewhere around 2000 kanji. But I'd guess those kanji won't be exclusively the jouyou kanji of RTK volume 1, since the 1503 kanji in Core5000 (the top 5000 words in Japanese according to latest Japanese corpus research) include some outside RTK1/3.

Also, according to this talk, averaged over various unspecified languages, 98% corresponds to 20'000 words, which is apparently how many words one learns up to adulthood (by year 20, people on average know 20'000 words, which they learn about 1000/year; another kanji learner made some notes at http://kazemakase.ca/2013/08/04/fluent-c...-polyglot/).

jmigno Wrote:Does this mean that 5000 words is far from enough to include most of the usual vocabulary, or that the official list of over 2000 kanji includes several hundred of unfrequent characters (somehow supporting the RTK-lite idea)?

Definitely the latter. The jouyou list contains many legacies of politicking and cultural posturing ("what will our ancestors think if we allow our children to never learn 藤!?" (藤 as in Fujiwara)). These 1503 kanji needed to "spell" the top 5000 words should take you to nearly 90% comprehension.

Edit: here's the 7.5 kanji in Core5k's words that aren't in Katsuo's combined-RTK1/3 spreadsheet: 炒贅籠餃姜壼埃 (and 々).