How Many Words Do You Need? Vocabulary Size and CEFR

"How many words do I need to know?" is one of the most common questions in language learning, and one of the hardest to answer neatly. Researchers have measured it carefully, but the answer depends on what you want to do, how a "word" is counted, and which kind of English you want to understand. This guide pulls together what the research says, and what it does not.

First, what counts as a word?

Counting words is harder than it sounds. Is decide one word, or do decides, decided and deciding count separately? What about decision and undecided? Researchers use three common units.

A token is any word you see in a text, counting repeats, so "the cat saw the cat" has five tokens. A lemma is a dictionary headword together with its inflected forms, so decide, decides, decided and deciding are one lemma. A word family goes further and groups a base word with its common derived forms, so decide, decision, decisive and undecided are one family.

Estimates in lemmas are always larger than estimates in word families, which is why you will see very different numbers quoted for the same thing. When you read any figure for vocabulary size, check which unit it uses.

How many words do you need to understand a text?

A well-known line of research asks a practical question: what share of the words in a text must you know to understand it without help? It is called text coverage. A frequently cited study built word-frequency lists from a large collection of British English and used them to estimate the answer. If a reader needs to know 98 percent of the words in a text for unassisted comprehension, then a vocabulary of about 8,000 to 9,000 word families is needed for written text, and about 6,000 to 7,000 for spoken text.

Another study with second-language learners looked at the relationship between vocabulary size, coverage and reading comprehension in more detail. It suggested two thresholds: an optimal one, about 8,000 word families giving 98 percent coverage, and a minimal one, about 4,000 to 5,000 word families giving 95 percent coverage. The study also found that even small increases in vocabulary knowledge contributed to better reading comprehension, even when they barely changed coverage.

These are research estimates and they depend on the text type, the word lists used and how comprehension was measured. They do not mean that you cannot read until you know 8,000 families. People read with less coverage and with help from dictionaries, pictures and context. But the numbers give you a useful sense of scale: comfortable, unassisted reading of ordinary English asks for a much larger vocabulary than conversation with a friend.

How many words do native speakers know?

It helps to compare yourself with native speakers, but the figures are easy to misread. One study that combined earlier research with a large online experiment estimated that an average 20-year-old native speaker of American English knows about 42,000 lemmas, which come from about 11,100 word families. The range was wide, from 27,000 lemmas for the lowest 5 percent to 52,000 for the highest 5 percent. The researchers noted that knowing a word can be shallow, as little as knowing that it exists, and that the very high numbers quoted elsewhere often include names and inflected forms.

The practical point is that native speakers do not know all of English either. They meet unfamiliar words all the time. A learner does not need to match a native speaker's total to communicate well.

Why the CEFR gives no number

You might expect the CEFR to say "B1 means 2,500 words". It does not. The framework describes what a learner can do, not how many words they know, and no official document lists the vocabulary of any level. What the CEFR is explains this in more detail.

Several projects map words to levels, and each one is an estimate built from data such as textbooks, exam papers and learner writing. They often disagree about individual words. Wordabulary does the same thing: each word in the library has a level from A1 to C2, and that tag is the app's own estimate of when a learner is likely to be ready for the word, not an official ruling.

Online lists that say "A2 means 1,500 words, B2 means 4,000" are therefore rough rules of thumb. They are most useful for setting expectations. You can expect your vocabulary to be a few thousand words at the middle of the scale and many thousand at the top, but you should not treat any single figure as a target.

Which words matters more than how many

Two people who both know 3,000 words can have very different practical ability, depending on which words they know. A small number of very common words cover a large share of everyday English. Learning those first gives the biggest return, and learning rare words before the common ones gives the smallest.

This is why vocabulary tools sort words by level and frequency. Starting with the most common words means every new word you learn shows up often, which also helps you remember it. The A2 guide and the B2 guide show what this looks like at two points on the scale.

Depth matters too. Knowing a word means more than recognizing it: you may also know how to pronounce it, what it goes with, and when it sounds too formal. A smaller set of words you can really use is worth more than a long list you only half recognize.

A rough way to check your own size

Several free online tests estimate vocabulary size by asking whether you know a sample of words and then scaling up. They are quick and can be motivating, but they carry the same caveats as the research above. The result depends on how the test counts words, the words in its sample, and whether it checks that you really know the meaning. Treat the number as a range, and notice which kinds of words you miss, since that tells you more than the total does. If you keep missing everyday words, work on those first. If you know the common words and miss the formal ones, that points to reading more demanding texts.

What to do with all this

Use the numbers to set expectations, and then let them go. Set goals in terms of what you want to do, such as following a team meeting, reading a news article or writing a clear email. Then build vocabulary in short regular sessions, starting with the words that appear most often at your level. How spaced repetition works explains why returning to words over several days beats a single long session.

If you want an estimate of where to begin, a placement quiz gives you a starting point, with the understanding that it is only a starting point.