Subtitles on, subtitles off: what watching TV really does for your vocabulary

Every language learner has told themselves the same comfortable story: that the hours spent on foreign box sets are not really procrastination but a kind of study by osmosis. It is the most appealing method there is, because it asks for nothing — no cards, no drills, just a sofa and a good series. The surprising thing is that the science does not laugh it out of the room. Watching television in a language you are learning genuinely teaches you words. What it will not do is quite what the fantasy promises, and the gap between the two is where all the useful advice lives.

Does the screen actually teach words?

It does, and someone has counted. In a 2018 study in Studies in Second Language Acquisition, Elke Peters and Stuart Webb sat learners down in front of a single full-length television documentary in their foreign language and, without warning, tested them afterwards on words that had appeared. The learners had picked up a real slice of the unknown vocabulary — on the order of one in seven of the words they hadn't known going in — purely as a by-product of watching. Nobody was studying. They were just following a programme, and some of the language stuck to them on the way past.

That single-sitting result sits inside a larger, calmer body of evidence. In 2013, Maribel Montero Perez, Wim Van Den Noortgate and Piet Desmet gathered the caption research into a meta-analysis in System and found a clear benefit of on-screen text for both vocabulary and listening. A much larger meta-analysis by Kanami Kurokawa in Language Learning in 2025 — pooling 89 samples from 49 studies — trimmed the estimate to a medium-sized effect rather than a spectacular one. That downgrade is worth trusting rather than mourning: it is what honest evidence looks like once the small, flattering early studies are outnumbered. The effect is real, it is repeatable, and it is moderate. Television teaches vocabulary the way a slow tide moves sand — not all at once, but genuinely, and more than you'd guess.

Why it works — and what does the actual work

The reason a screen can teach a word at all is that it does several things a flashcard cannot do at once. The word arrives inside a real sentence, spoken by a person with a face, in a situation you can see. Meaning, sound and image land together, which is close to the ideal way for a new word to be filed. It is also, quietly, the same word said by many different characters, in many voices and moods — the very variety of voices that keeps a word from getting welded to a single recording.

But the engine underneath all of it is plainer, and Peters and Webb put their finger on it: the words that stuck were the words that recurred. The single strongest predictor of whether a learner picked up a word was how many times it turned up on screen. This is the same unglamorous truth that governs vocabulary everywhere — that you don't learn a word the first time you meet it, but somewhere around the tenth or fifteenth, spaced out and dressed in slightly different clothes each time. Television is a machine for delivering exactly that: repeated, incidental, mildly varied encounters. A murder mystery will say evidence and suspect forty times before the credits roll, and those are the words you leave with.

You don't learn the words a show uses. You learn the words a show keeps using.

The subtitle question

Which brings us to the argument every learner eventually has with themselves: subtitles on or off, and in which language? Here the research is honest enough to be a little unsatisfying. Some captions almost always beat no captions — on-screen text gives the ear something to check itself against, and turns a blur of sound into words with edges. Beyond that, there is a trade-off rather than a winner.

Subtitles in your own language are the comprehension crutch: they let you follow a plot that would otherwise lose you, and comfort is not nothing, because a show you can bear to finish teaches more than a masterpiece you abandon. But they come with a catch every honest viewer already knows — the eye goes to the language it reads fastest, and you can watch an entire season in translation while barely hearing the foreign words at all. Captions in the target language ask more of you and, in return, do something the translation can't: they tie the sound you're hearing to the written shape of the word, which is why the caption studies tend to find their firmest gains in learners recognising and spelling the word forms afterwards. The rough rule that falls out of all this is less about a correct setting than about attention. Whatever keeps you actually noticing the foreign language — rather than reading past it — is the setting that teaches.

The honest catch: you need a doorway in

Now the part the fantasy leaves out. All of the above assumes you can already understand most of what you're hearing, and for a true beginner that assumption is false. Webb and Rodgers, in a careful 2009 study in Language Learning, worked out just how much vocabulary ordinary television demands. Knowing the most frequent 3,000 word families (plus names and the odd ah and hmm) covered about 95% of the words in the programmes they analysed; reaching 98% took around 7,000. To follow a given genre at that comfortable 95% mark generally took somewhere between 2,000 and 4,000 word families.

That threshold is the whole story of why TV disappoints so many beginners. Below it, the dialogue isn't a stream of mostly-known words with a few new ones bobbing in it — it's a wall, and a wall teaches nothing, because a word you can't get any purchase on can't attach to anything you know. This is the same reason the first couple of thousand words matter so disproportionately. They are the doorway. Cross it, and television turns from noise into one of the richest vocabulary sources you'll ever have. Stand outside it, and the most beautifully subtitled series in the world will wash straight over you.

What to actually do

If you're past that doorway, television becomes a serious tool rather than a guilty pleasure — provided you use it a little deliberately:

• Pick something slightly below your level. You want a show where most words already land and only a few are new. Comprehension isn't the enemy of learning here; it's the precondition for it.
• Watch narrowly, not widely. Sticking with one series, or one genre, means its particular vocabulary keeps recurring — and recurrence is the engine. A single detective show teaches its own words far better than ten unrelated films.
• Use captions in the language you're learning, once you can. They bind the sound to the written form. Fall back to your own language only when the plot would otherwise slip away entirely.
• Rewatch without guilt. The second pass is where words you merely glimpsed the first time get their extra encounters — the ones that move them from familiar to yours.
• Catch the words that matter to you. When a word keeps returning and you still don't know it, that's the show telling you it's worth learning properly. Note it, and give it a home outside the screen.

Where vocabla fits

Television's great weakness is the flip side of its charm: it is generous but random. The word you most need might surface once, in a mumble, and never come back; the word you already own gets said a hundred times. Left to the screen alone, you keep whatever the plot happened to repeat and lose whatever it didn't — and the useful word that appeared just once is exactly the one that slips away by morning.

That's the gap vocabla is built to close. Watch the show for the rich, real, in-context exposure only a story gives — then hand the words worth keeping to something that won't let them fade. The app turns a one-off encounter into the spaced, deliberate returns that actually fix a word in place, meeting it again inside a sentence and asking you to recall it just before you would have forgotten. Television gives you the words; vocabla makes sure the good ones stay. Watch on — and let the screen be the beginning of learning a word, not the whole of it.