Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The most frequent 777 characters give 90% coverage of Kanji in the wild

japanesecomplete.com

Post reply on HN
Print view View original

41–50 of 214 posts

Re: The most frequent 777 characters give 90% coverage of Kanji in the wild

#41

Post by sova »

quicklime wrote: post #25
ekianjo wrote: post #14

Lol, 777 kanji is just too short to even be comfortable reading emails at work in Japanese. So what good is 90 percent if the only thing it allows you is to do shopping?

777 of the most popular kanji from general Japanese might not be useful for work emails, but if you find the top 777 from a corpus of your own emails, it might be. It's not hard - I've done this before using a corpus made up of work documents. There's a part-of-speech analysts tool called mecab that gives the word stems, and makes it easy to find word boundaries (since Japanese doesn't use spaces). The output went in…

genius

Re: The most frequent 777 characters give 90% coverage of Kanji in the wild

#42

Post by bdowling »

jacobolus wrote: post #19
wrp wrote: post #7

Vocabulary frequency lists have long been very popular with publishers, so I assume as well with language learners. The assumption behind these is that with knowledge of, say, 90% of the vocabulary you encounter, you will comprehend 90% of the writings you encounter. It doesn't work out at all that way, though, because it is generally the low frequency vocabulary that carries the key information in a text and on whic…

In the US, there are beginning reader books (e.g. the “I Can Read!” and “Ready to Read” series) which intentionally use somewhat limited vocabulary. These are somewhere between a picture book and a chapter book: they have usually 1.5 pages of text and 0.5 pages of picture in each 2-page spread; the text is set in a large font but there are at least a few sentences per page; usually the books are 30–50 pages long, wit…

Notable examples are The Cat in the Hat, which uses 236 words, and Green Eggs and Ham, which uses only 50 words. (Wikipedia)

Re: The most frequent 777 characters give 90% coverage of Kanji in the wild

#43

Post by mantap »

wrp wrote: post #7

Vocabulary frequency lists have long been very popular with publishers, so I assume as well with language learners. The assumption behind these is that with knowledge of, say, 90% of the vocabulary you encounter, you will comprehend 90% of the writings you encounter. It doesn't work out at all that way, though, because it is generally the low frequency vocabulary that carries the key information in a text and on whic…

When you want to learn an abstract word, you have to learn it in the context of a sentence in the foreign language. In order to do that you have to know the meanings of the other words in the sentence.

This works passively as well, if you see a sign and you understand 90% of what's written on it but there's one word you don't understand, you will much more easily remember that word later because you can place it in context.

Re: The most frequent 777 characters give 90% coverage of Kanji in the wild

#44

Post by wilsonthewhale »

jackklika wrote: post #27
dvduval wrote: post #4

Yes, I have been studying Chinese for about five years now and I can recognize about 2700 characters. Looking at this Japanese character set, I'm guessing I know 80% or so. I always act like charts like this.

I've noticed the same where I can understand the basics of a Japanese program if there's subtitles. But there are some oddities like 食 being used in a verb (as in 食べる) or 行 as to go. But you still have good context and can at least understand the topic in most cases.

knowing Cantonese helps here, as both the verbs you mention are still common verbs in modern Cantonese.

Re: The most frequent 777 characters give 90% coverage of Kanji in the wild

#45

Post by H8crilA »

sova wrote: post #40
H8crilA wrote: post #38

Yet another (re)discovery of Pareto's distribution. Net worth, stock returns, popularity of words in languages, casualties in wars or natural disasters, size of cities, popularity of artwork pieces like songs or computer games, ... https://en.wikipedia.org/wiki/Pareto_distribution#Applicatio...

This is closer to 90:25 rather than the typical Pareto of 80:20

80:20 is a pop-culture take on the distribution. It does a good job of visualising it to someone who doesn't know maths.

It's really any distribution with the CDF of the form x^(-a)

Re: The most frequent 777 characters give 90% coverage of Kanji in the wild

#46

Post by bgee »

redthrow wrote: post #36

If Japanese people just used kana (like Korean people use hangul today), kids in Japan don't have to spend countless hours learning this complex (and often irrational) writing system. http://kanamozi.org/hikari959-04.html > If you really want to be native-level Japanese, kanji are essential There are visually impaired people who have difficulty learning kanji but speak Japanese fluently. Language is not just for peop…

The English language is also quite irregular both in spelling and grammer[sic]. Maybe we should start with it given that children around the world "spend countless hours learning this complex (and often irrational) writing system".

Disclaimer: native Chinese speaker, non-fluent Japanese speaker, English sufferer

edit: disclaimer and grammar...

Re: The most frequent 777 characters give 90% coverage of Kanji in the wild

#47

Post by matt-attack »

kazinator wrote: post #33

I know around 10,000 Japanese words. As far as kanji goes, I've studied some 1200 of them intensively; through vocab I know many more as parts of words. I quite often have to reach for a dictionary when reading. If you know 777 kanji in some way (like associating them with meanings, through your native language) and you haven't crammed on any vocabulary, you absolutely will not be able to read a thing. In fact, even…

> I quite often have to reach for a dictionary when reading

So when reading kanji, how do you look up a word (picture) you don't know? Since there isn't a minimal set of characters, the notion of "alphabetical order" seems impossible. Weird that I've never thought about this until now, but I'm honestly baffled.

Re: The most frequent 777 characters give 90% coverage of Kanji in the wild

#49

Post by echelon »

Does anyone have frequency lists of vocabulary broken down by type (verb, noun, etc.)? I've seen word lists on Wiktionary, and I'm attempting to cross-reference jisho and other sources.

How many words do you need to comprehend for daily competency? Would the 10,000 suffice?

How many words do you need to be able to watch anime aimed at children (eg Bono Bono) or teenagers (Boku no Hero Academia)?

Re: The most frequent 777 characters give 90% coverage of Kanji in the wild

#50

Post by redthrow »

bgee wrote: post #46
redthrow wrote: post #36

If Japanese people just used kana (like Korean people use hangul today), kids in Japan don't have to spend countless hours learning this complex (and often irrational) writing system. http://kanamozi.org/hikari959-04.html > If you really want to be native-level Japanese, kanji are essential There are visually impaired people who have difficulty learning kanji but speak Japanese fluently. Language is not just for peop…

The English language is also quite irregular both in spelling and grammer[sic]. Maybe we should start with it given that children around the world "spend countless hours learning this complex (and often irrational) writing system". Disclaimer: native Chinese speaker, non-fluent Japanese speaker, English sufferer edit: disclaimer and grammar...

I'm a native Japanese speaker but as far as the writing system goes, I fully agree with the linguist Geoffrey Pullum's following statement about Chinese characters / kanji:

In consequence, this horror-show of a writing system, with its crippling memorization burden for students and malign impediment to progress in science and industry, is the focus of so much intellectual investment and cultural pride that getting rid of it is out of the question. Intolerable though it is, it will continue to be tolerated – leaving English, with a spelling system that positively stinks, smelling almost like a rose.

https://www.chronicle.com/blogs/linguafranca/2016/01/20/the-...

Don't get me wrong, I'm generally in favor of made-in-China products, but those characters to me were the worst made-in-China product I wasted so much time on (and unlike video games, I wasn't even having fun with them.)

Post reply on HN