How many words do you need to read the news in another language?
Research says you need to know about 95 to 98 percent of the words on a page to read it comfortably. What that means in word counts, and how to get there.
The short answer: to read news comfortably without help, you need to know about 98 percent of the words on the page. In English that takes roughly 8,000 to 9,000 word families. With about 95 percent, around 4,000 to 5,000 families, you can follow the main points, and the rest is guessing. You can start reading real news well before either number if something explains the unknown words for you as you go.
Why the percentage matters more than the total
Researchers measure readability as coverage: the share of words in a text that the reader already knows. Two studies set the figures most teachers use.
- Hu and Nation (2000) gave learners the same story with different shares of unknown words. Comprehension rose steadily with coverage, and most readers only understood the text well at around 98 percent.
- Laufer and Ravenhorst-Kalovski (2010) proposed two thresholds: a minimal one at 95 percent coverage and an optimal one at 98 percent.
98 percent sounds high. It still means one unknown word in every 50, or about one in every two or three lines of a newspaper column. At 95 percent it is one in 20, more than one per line, and reading turns into decoding.
How many words that is
Paul Nation worked out how many word families give those coverages in real English text. His 2006 estimate is that 8,000 to 9,000 word families cover 98 percent of written text such as newspapers and novels, and 6,000 to 7,000 cover 98 percent of spoken English. Laufer and Ravenhorst-Kalovski put 95 percent at about 4,000 to 5,000 families.
A word family groups a word with its forms: decide, decides, decided, decision, decisive count once. So the number of separate words you know is higher than these figures.
Most research is on English. Languages with lots of inflection, like Turkish, Finnish or Russian, have more forms per family, and languages that share many words with yours, like Spanish for an English speaker, can feel easier than the numbers suggest.
What this means for a learner
- Course vocabulary is not enough. A full beginner course covers maybe 1,000 to 2,000 words. That is well under 90 percent coverage of a news article, which is why the first real article feels like a wall.
- Easy news is a real bridge. Outlets written for learners keep vocabulary at a lower level: RFI Journal en français facile, nachrichtenleicht and DW’s slowly spoken news in German, 8 Sidor in Swedish, NOS Jeugdjournaal in Dutch, NHK News Web Easy in Japanese.
- The last words come from reading itself. The words between 3,000 and 9,000 are each fairly rare. You meet them by reading a lot, and you keep them by seeing them again.
Reading the news before you reach 98 percent
You do not need to wait. What makes a text with too many unknown words frustrating is the cost of each lookup: leaving the page, choosing between dictionary meanings, losing the sentence. If the cost of a lookup is close to zero, a text at 90 to 95 percent becomes readable.
That is the idea behind ContextCat. Today brings headlines from real outlets in the language you read. Tap a word you do not know and a note under the sentence says what it means here, in your language. The word comes back later as a card, so your coverage creeps up with every article.
A simple plan
- Read one short article a day, at the edge of what you can follow.
- Look up any word that stops you understanding the sentence. Skip the ones you can guess.
- Review the words you looked up, in their sentences.
- Every few weeks, try an outlet one step harder.
Questions
How many words do I need to read a newspaper?
- For English, studies estimate about 8,000 to 9,000 word families to know 98 percent of the words in written text, which is when reading becomes comfortable without help. Around 4,000 to 5,000 word families gives about 95 percent, enough for a basic understanding.
What is a word family?
- A base word with its forms and close relatives. Read, reads, reading, reader and readable count as one family. That is why the numbers are smaller than counts of every distinct word.
Can I read the news before I know that many words?
- Yes, with help. A tool that explains the unknown words in place, or easy-news outlets written for learners, lets you read real news far earlier. Each word you look up and see again moves you toward the threshold.
Sources
- Hu and Nation (2000), Unknown vocabulary density and reading comprehension, Reading in a Foreign Language 13(1)
- Nation (2006), How large a vocabulary is needed for reading and listening? Canadian Modern Language Review 63(1)
- Laufer and Ravenhorst-Kalovski (2010), Lexical threshold revisited, Reading in a Foreign Language 22(1)