Everyone learning English has asked this, usually in the middle of a movie, after getting stuck on the fifth unknown word in two minutes.
The good news is that there is an answer, it comes from corpus linguistics research, and the number is much smaller than it seems. The even better news is that the order in which you learn words matters more than the total.
The numbers
Research on language comprehension uses a measure called coverage: the percentage of words in a text or in speech that you know. The most commonly cited thresholds:
| Words you know | Coverage of everyday speech | How it feels |
|---|---|---|
| 1,000 most frequent | About 80% | You get the topic, miss the meaning |
| 2,000 most frequent | About 90% | You follow with constant effort |
| 3,000 most frequent | About 95% | You follow comfortably, stuck now and then |
| 5,000 most frequent | About 98% | Rarely stuck, you guess from context |
| 8,000 to 9,000 | Close to 99% | Adult native-like comprehension |
The number that tends to surprise people: 95% coverage still means one unknown word in every twenty. In movie dialogue, that is roughly one word every two lines. It sounds small, but it is enough to miss the joke.
And the encouraging number: the difference between “I miss everything” and “I follow well” lies between 2,000 and 3,000 words. Not 20,000. Three thousand.
Why frequency is everything
English, like every language, is wildly uneven in how words are distributed. The 100 most common words cover about half of everything people say. The 1,000 most common cover 80%. After that, the return drops fast: going from 95% to 98% takes about two thousand more words, and going from 98% to 99% takes three thousand more.
In practice, that means your first thousand words are worth more than the next five thousand. And it is why memorizing lists of “advanced vocabulary” while the basics are still missing is a bad deal.
Knowing a word is not memorizing it
Here is the problem with the number on its own. “Knowing a word” has levels, and only the first one is quick:
- Recognizing it when reading. You see it written and remember. This is the easiest level and the one word lists give you.
- Recognizing it when listening. You hear it and identify it, at normal speed, with an accent, inside a sentence. That is a huge jump from the previous level, and it is where almost everyone gets stuck with movies.
- Producing it. The word comes to you on its own when you need it. This level takes years and only comes with use.
A typical learner “knows” 4,000 words at level 1 and maybe 1,500 at level 2. That is exactly why they can read an article easily and still not follow a scene from a series.
The practical conclusion: you do not need more words, you need the same words in your ears. And that is only solved by listening, a lot, with subtitles as support. More on training that in how to improve your English listening.
The phrasal verb problem
There is one category that breaks the math and deserves its own section. Get up, get over, get along, get by, get through: you know “get” and you know “over”, and you still have no idea what the combination means.
Spoken English is full of these. Reasonable estimates put a few hundred phrasal verbs as essential for everyday conversation, and they are especially cruel because each one can have three or four meanings depending on context.
You cannot deduce them and memorizing lists does not help. The only thing that works is meeting the same phrasal verb in different situations, several times, until the meaning separates itself in your head. Watching is the most efficient way to get that repetition. A starting set is in the 35 most common phrasal verbs.
And false friends
Another trap the coverage numbers hide: words you think you know. A Spanish speaker “knows” actually because it looks like actualmente; a German speaker “knows” become because it looks like bekommen. Both are wrong, and both feel right.
What makes these words dangerous is that you do not know you got them wrong. An unknown word sets off an alarm and you look it up. A false friend slides by, you understand the sentence wrong and move on confidently. We collected the main ones for each language in false friends in English.
What to do with these numbers
If you are below 2,000 words: prioritize frequency. Simple, repetitive content with limited vocabulary: children’s series, classic sitcoms, educational channels. Nothing to be embarrassed about: it is the shortest path. See the list of series by level.
If you are between 2,000 and 4,000: your bottleneck is almost certainly not vocabulary, it is your ear. You understand reading and not listening. The fix is a high volume of audio with dual subtitles, so every word you already know in writing gains its sound version too.
If you are above 4,000: now it is worth hunting specific vocabulary, phrasal verbs and idioms. And it is worth gradually dropping the translation subtitle.
A calculation that helps you decide
Watching 30 minutes a day with dual subtitles and looking up what you do not understand, a person meets between 10 and 30 new words per session. Not all of them stick. A realistic retention rate for a word seen in context and looked up on the spot is somewhere between 15% and 25%.
That works out to 3 to 7 words learned per day. It sounds like little. In a year, it is more than fifteen hundred words, at level 2, with sound and context attached.
Fifteen hundred words a year, starting from 1,500, puts you at 3,000 in one year and 4,500 in two. That is exactly the range where a movie stops being an exercise and becomes a movie.
It is not fast. But it is what works, and it is a lot faster than the course you started and dropped three times. Looking words up without leaving the video, as Telidioma lets you do on Netflix and YouTube, is what makes those 10 to 30 lookups per session realistic instead of exhausting.