How many words do you need to watch TV without subtitles?

"How many words do I need to know to watch without subtitles?" It sounds like a question with no answer, but applied linguistics has been studying exactly this for decades, analyzing the scripts of hundreds of movies and TV programs. The answer isn't a magic number, but it's a lot more concrete than "a lot."

How many words do you need to understand a show without subtitles?

Somewhere between 3,000 and 7,000 word families, depending on how much you want to understand. The main studies:

In both of Webb and Rodgers's studies, the numbers assume you also know proper nouns and interjections (like oh, um, uh). Ha's study used newer frequency lists, which combine British and American English, and a much larger body of data, which helps explain the difference at the 98% level.

What do "word family" and "95% coverage" mean?

They're the two units the research uses, and understanding both keeps you from drawing the wrong conclusions.

A word family is a base word plus its inflected and derived forms. Decide, decided, deciding, decision, undecided: all of it counts as one family. So "3,000 families" is far more than 3,000 forms you need to recognize.

Coverage is the percentage of the spoken words that you know. In practice:

• 95% coverage = one unknown word in every 20.
• 98% coverage = one unknown word in every 50.

The difference looks small on paper and is huge on screen. In the first case you stumble all the time; in the second, every now and then. (What that means in words per minute is in The words your English is missing.)

95% or 98%: how much counts as "understanding"?

It depends on what you call understanding. Researchers usually treat 95% as the threshold for reasonable comprehension and 98% as the threshold for comfortable comprehension.

Listening comes with its own debate. The 98% threshold came from reading studies. Hilde van Zeeland and Norbert Schmitt tested comprehension of spoken narratives and concluded that, for listening, 95% already gives fairly good comprehension, which would take 2,000 to 3,000 families. Paul Nation, on the other hand, points out that speech goes by and doesn't come back, which may call for even more than 98% to follow it comfortably. The picture helps: in movies and shows, what you see makes up for part of what you don't understand.

If you want to get past "I sort of understood" and reach the point of understanding every line without effort, the realistic target is the 98% range.

Why do the numbers vary so much?

Because every show is its own case. Even within the studies themselves, the variation is large:

• Genre. For movies, reaching 98% took 5,000 to 10,000 families, depending on the genre. For TV, 5,000 to 9,000.
• Origin. American movies reached 98% with 6,000 families; British ones, with 7,000.
• Format. Ha found soap operas less demanding than movies and TV programs.
• Episode. Webb and Rodgers observed wide variation in coverage from one episode to the next.

So be wary of any promise along the lines of "learn X words and understand everything." The numbers are an order of magnitude, not a finish line.

Why do the most frequent words pay off the most?

Because a small number of words accounts for most of what people say. Nation analyzed the script of Shrek and measured how much of the movie each band of a thousand families covers, including proper nouns:

The first thousand families cover more than four-fifths of the movie. The sixth thousand adds less than half a point. (The jump at the third band is inflated by interjections like oh, um and er, which fall into that band of the list used.) In spontaneous conversations analyzed in the same study, the 2,000 most frequent families already covered about 89% of the words.

That has two consequences. First: studying in order of frequency is the most efficient path. Second: after the basics, each new word adds less coverage, which is why intermediate learners feel like they study and go nowhere. The last few percentage points are made of thousands of words that rarely come up, but they're what separates "sort of" from "got all of it."

So what should you study first?

Three criteria, in this order.

1. Frequent words you still don't own in speech. The lists Nation used were built from a corpus that was 90% written, and he himself notes that lists based only on speech would put several words in more frequent bands. Here's where some words from Shrek fall on that list:

None of them is rare on an American show. In other words, a word can sit "far down" the general ranking and still be common in the spoken English you want to understand.

2. The topic words of your show. In Shrek, words that rank low in general frequency, like ogre, swamp and donkey, show up dozens of times. Nation tested the effect of treating the most repeated ones as already known: the movie reached 98% coverage with a thousand fewer families. He himself notes that the gain in coverage is small, but the practical point stands: a word that keeps coming up in what you watch pays off more than its ranking suggests. If you watch a hospital, courtroom or police show, the topic vocabulary will come back every episode. It's worth learning early.

3. Everything else, in order of usefulness. Low-frequency words that appear once and never come back can wait.

dailyBe helps with what comes after that decision. In the AI dictionary, you look up the word, expression or phrasal verb you heard on the show and see at least five example sentences; the ones that make sense to you go into your list and become reviews. You decide what goes in, using the criteria above: frequent in speech, repeated in your show, or important for your work. From there, the app tracks each word by how many times you get it right compared with how many times it has come up, and once a word is learned, it leaves the spotlight to make room for the next ones. See Features.

Where to start?

Pick one show you want to watch without subtitles and stick with it for a few episodes instead of jumping from one to another: that way the topic vocabulary repeats in your favor. Write down the sentences you didn't understand, check whether the words are frequent, and start with those. And since a word you learn and don't review tends to fade, pair this with spaced reviews, explained in Spaced repetition: what it is and how it works.