# Number of unique characters in the X most common words?

I assume this would be wouldn't be too hard to calculate with a proper database, but does anyone know the number of unique characters/morphemes used in the 5000, 10000, 30000, etc. most common words?

For example, I heard that 6000 characters covers over 99% of words in modern usage, and assuming that a native's passive vocabulary is around 30000, I conclude that there are approximately 6000 unique characters used in the 30000 most common words (this could be way off).

I am curious about these statistics. Preferably, someone has a nice plot showing the relationship between the two, but a few data points would also be appreciated.

• Are the "words" in your question 2-character or 3-character phrase like "因為", "知道", and so on? Jul 15 '14 at 1:45
• It seems what you need is 字频表. This is one example, you can find 3591 characters have already covered 99% of words in that corpus. And for this example (traditional characters), 3047 characters for 99% coverage.
– Stan
Jul 15 '14 at 2:45

I write a simple script to compute the plot from an old phrase frequency data.

If you're interested, here is the raw data.

These data are for Taiwan's traditional Chinese,
but the statistics should be very similar to simplified Chinese.

I'm curious about that "30,000 words for a native" number. HSK lvl 5 at 2,500 words and 1,685 characters is what allows you to read a newspaper. HSK lvl 6, at 5,000 words, is regarded as complete fluency of the Chinese language. There will obviously be people who know many more words, but I don't think when 5,000 words means fluency, that a native would know 30,000, that's a massive amount of words. The literacy rate in China is at 95%, which is still relatively low, I don't think a native knows 30,000 words, maybe around 10,000 if I had to make a wild guess. According to a harvard study the average English speaker knows 12,000 words.

• This, first off is totally not an answer and would be more suited as a comment. Secondly, literacy has nothing to do with the amount of words somebody knows (granted there is a bit difference between formal and informal words in Chinese but even formal words would be used on tv and on the radio). Thirdly HSK is used to measure proficiency and not fluency - someone at level 6 has ok Chinese but I wouldn't venture to say it's anywhere near fluent.
– Mou某
Jul 15 '14 at 14:20
• Education is obviously linked to the amount of characters and words you know, GLM and Harvad showed that high school students have a smaller vocabulary than college students. For High school students their vocabulary was around 10,000, some college students have a vocabulary of 20,000. However, I think the 30,000 number is a bit high for natives in China. Jul 15 '14 at 14:25
• OP is talking about the number of words people know - I'm sure the illiterate know a number of words without knowing characters...
– Mou某
Jul 15 '14 at 14:34
• As someone who is at HSK6 level I can assure you it is absolutely nowhere near "complete fluency". When I watch Chinese TV I understand less than half of it. Apr 25 '18 at 8:41