How many Chinese characters do you need? We measured it (video)
Unicode lists 101,997 Chinese characters, but you need far fewer. We measured how much 1,000 to 3,000 characters let you read: signs, menus, news and novels.
Chapters
- 0:00Start
- 0:19The short answer
- 0:36How many characters are there?
- 1:14A thousand characters do most of the work
- 2:17Every goal needs a different number
- 3:08Menus and signs reward a short list
- 3:52What each HSK level lets you read
- 4:35Chinese children learn about 3,500 characters
- 5:18Knowing the characters isn't knowing the words
- 5:58A plan, and what to do next
Transcript
Start
Unicode now lists 101,997 Chinese characters. But just a thousand of them cover 94% of the characters in film and TV subtitles. From SayMei Guides, by Mei Lin: How many Chinese characters do you need? We measured it.
The short answer
The short answer: about 100,000 characters exist, but news, TV and novels run on about 2,500 of them. Menus are the exception. They're full of food characters you rarely see anywhere else, so it pays to learn those on purpose.
How many characters are there?
So how many are there? We counted Unicode's newest release: 101,997 Chinese characters. The biggest printed dictionary, the Zhonghua Zihai, lists more than 85,000. Most of these are old forms, variant spellings, or one-off names of people and places. Now compare that with real use. 193 million characters of modern writing use fewer than 10,000 different ones. China's standard list has 8,105. That's the pool you're really drawing from, and you only need the top of it.
A thousand characters do most of the work
In film and TV subtitles, the most common one, 我, I, is about one character in every 23 that people say on screen. Learn the top 250, and you know 74% of the characters in film and TV. The top 1,000 cover 94%. News is harder: 1,000 characters cover 85% of it, and 2,000 cover 96.5%. Percentages hide how reading feels. At 90%, one character in 10 is new. That's study, not reading. At 95%, one in 20: you can read with a dictionary close by. At 98%, one in 50, and you can usually guess the rest from context. The guide has a calculator for this: pick your number or your HSK level, and a goal, and it blanks out every character you'd still have to look up. Some never turn up on any HSK list, like the second character of 悉尼, Sydney, so they stay blank at every level.
Every goal needs a different number
So how many do you need? It depends on what you want to read. To know 95 to 98% of the characters, TV and film take about 1,100 to 1,700. News, 1,800 to 2,400. Novels, 1,900 to 2,700. And restaurant menus need 3,800 to 4,600. Why so different? Everyday talk keeps repeating the same characters. News is full of names and places, like 湾, bay, as in Taiwan. Signs use a few characters you rarely see elsewhere, like 勿, don't, as in 请勿吸烟, please don't smoke. And menus are packed with ingredients and cooking methods. If you have one goal, aim at its number.
Menus and signs reward a short list
Menus show this best. In everyday order, even 2,000 characters leave you knowing under three quarters of a menu. The gaps are the food itself. 虾, shrimp. 鸭, duck. 蒸, steam. 蟹, crab. The single biggest gap is 炒, stir-fry: the most common menu character that everyday order leaves until after the first thousand. So we tested a shortcut. We built a list from half the dishes in Beijing's official menu guide and checked it against the other half. To know 90% of a menu's characters, everyday order takes 2,808. A list made for menus takes about 355.
What each HSK level lets you read
What about HSK? The new HSK has two character lists: one to recognise, and a shorter one to write. Many pages still quote the old exam's HSK 6 figure of about 2,600 characters. The new reading list for levels 1 to 6 has 1,940. Here's how much of the news each level's list covers. HSK 1, with 246 characters: 36%. HSK 4: 81%. HSK 6: about 95%. The top band, HSK 7 to 9, has 3,088 characters and reaches 98%.
Chinese children learn about 3,500 characters
How does that compare with Chinese children? China's 2022 curriculum expects pupils to recognise about 1,600 characters by the end of grade 2, and 3,500 by grade 9. They learn to write fewer. So a Chinese second-grader is expected to recognise about as many characters as the HSK 5 list. The legal bar for adult literacy is lower. A State Council rule, first published in 1988, counts town residents as literate at 2,000 characters, and farmers at 1,500. You'll often read that educated Chinese adults know 8,000. We couldn't find a study behind that number, so we don't repeat it as fact.
Knowing the characters isn't knowing the words
Here's the catch: you read in words. Three in four words on the HSK 1 to 6 lists are two characters long, and the pair often means something neither half suggests. 东, east. 西, west. Put them together, and 东西 means thing. Take the news. With HSK 6's characters, you'd know about 95% of the characters. But only about 62% of the words are on the HSK 1 to 6 word list. Almost a third are made of characters on that list, yet aren't words on it. Those are the words you can sound out but not understand.
A plan, and what to do next
So you don't need 3,000 characters before you start reading. You need the right next few hundred, met many times. One: pick one goal. Two: learn characters inside words. Three: read a lot of easy text. Four: review on a schedule. Five: write less than you read. By HSK 6, you recognise 1,940 characters but write only 700. What to do next? Check your number in the calculator, with the goal you care about most. For menus or signs, learn that goal's short list first. And read something easy today: one short graded story beats an hour of character lists. And chapter 1 of every SayMei story is free. And to practise speaking, you can talk with Mei Lin for free: your first lesson is 10 minutes, with no account and no card. The full guide, with the calculator, is on the page at saymei dot app.