Almost everyone knows this person, and in several large European countries this person is the majority. They studied English for eight years and passed the exam. Ten years later they can read a menu, catch perhaps a third of a news broadcast, and cannot get through five minutes of conversation without apologising for their English.

The usual explanation is that school language teaching is stuck in the nineteenth century. That is backwards. The teaching has changed, over and over. Something else is holding the results down, and it can be measured.

The short version

  • The method has been replaced four or five times since 1882. The results barely moved.
  • In a class of 28, a pupil speaks for about 43 seconds a lesson. Six years of school comes to roughly eight hours of talking.
  • School hands a pupil around 700 hours of language, which on paper should reach C1. Almost nobody gets close, because most of those hours are not spent using the language.
  • Europe's biggest study of school language learning did not test speaking. It was too hard to organise: the whole problem in one line.
  • The countries that do better did not find a better method. They found more hours: other subjects taught in the language, television subtitled instead of dubbed, a daily life that needs the language.

The method has been thrown out four times, and nothing moved

Grammar-translation, the method everyone pictures when they complain, was already being attacked as a scandal in 1882, when a German philologist published a pamphlet whose title was its whole argument: Der Sprachunterricht muss umkehren! Language teaching must turn around. Then came the direct method, then drills and the language laboratory, then communicative teaching, then the Common European Framework, then teaching other subjects through the foreign language. Four or five complete overhauls in 140 years, each sold as the end of the old way.

Then the results, measured properly for the first time in 2011. Europe tested almost 54,000 pupils across 16 school systems. In their first foreign language, 42% reached B1 or B2, the level where you can hold your own. Exactly the same share, 42%, were at A1 or below, which means barely started. Malta and Sweden had 82% of pupils at B1 or above in English. France had 14%. England had 9% in French.

The method changed and the outcome did not, which means the method was never what was holding things back.

Six years of school buys about eight hours of speaking

Cambridge English publishes the hours its own exams assume: roughly 200 per level, 700 to 800 to reach C1. Now count what school hands over. Three lessons a week of 45 minutes, 38 weeks a year, is about 85 hours. Six years of that, plus a few hundred hours in primary, comes to roughly 700. On paper, C1.

Almost nobody reaches C1, so those are not real hours. Some go to settling the class and handing back work. Some go to the mother tongue, which is where instructions are given and grammar is explained. Some go to writing, which is silent. What is left, the part where a pupil actually says something in the language, is shared out among everyone in the room.

Give a very good teacher 20 minutes of pupil speech in a 45-minute lesson. In a class of 28 that is 43 seconds each. A school year is about 80 minutes. Six years is roughly eight hours.

Eight hours of talking in six years, and then we are surprised that nobody can talk.

Communicative teaching found the one way out of that arithmetic decades ago: put the class in pairs. Twenty minutes in pairs gives everyone ten minutes instead of 43 seconds, fourteen times as much. That is precisely why the reform happened, and precisely what large classes, noise, fixed rows of desks, mixed ability and an exam that ignores speech all work against. The method knew the answer. The classroom made it almost impossible to use.

Europe's biggest language study skipped speaking because it was awkward

That 2011 survey tested reading, listening and writing. It did not test speaking, which the report itself called logistically difficult for a first round. The most careful measurement Europe has ever attempted left out the one skill the entire public complaint is about, because it was hard to organise.

That is a symptom, and it is the most useful fact in this article. Exams decide what gets taught. Teachers teach towards the exam, because that is what their pupils need and what their school is judged on. If the exam is a reading comprehension, a grammar section and a written text, then the exam has defined the subject, and talking is not part of it.

Japan tried to change exactly this and failed. Commercial proficiency tests were to be folded into the national university entrance exam, purely to get a speaking section into it, and the plan collapsed on fairness grounds: who lives near a test centre, and who can pay the preparation industry around those tests. Speaking exams are not impossible. Handing them to a commercial market is what kills them.

Japan is fifth in the world at maths and near the bottom at English

Japan scored 536 in the 2022 international school maths tests against a rich-country average of 472, and sits near the top in reading and science too. In the best known English proficiency ranking for 2025 it came 96th out of 123 countries, in the lowest band. That ranking is built from people who choose to take a free online test, so treat the exact place with care, but nobody in Japan argues with the direction.

A system that teaches maths better than almost anywhere cannot teach English to the middle of the pack. It is not aptitude and it is not effort. It is four things:

  • Distance. Japanese takes an English speaker three to four times the hours French does, and the difficulty runs both ways.
  • Exams. Japanese entrance exams have long rewarded reading, grammar and translation, which can be marked at scale by people who are not fluent speakers themselves.
  • Exposure. Japan dubs its television, and English barely shows up in an ordinary day. A Dutch teenager meets English undubbed on screen, in games, in music, in half the internet they use, and later in a work email.
  • Need. You can live an entire professional life in Japan without English. In the Netherlands you cannot, and every fourteen-year-old knows it.

Which countries dub and which subtitle was settled in the 1930s on market size and politics, not education. Economists have used that accident as a natural experiment, and the result is one of the clearest findings in the whole field: subtitling countries score markedly higher in English, and the gap is widest in listening. Several hundred hours of listening a year, at no cost to any ministry.

France, at 14%, is the check on any smug reading of all this: it dubs too, and it does not need English to get through the day. This is not national character. It is how many hours of a language a country's ordinary life gives away for free.

Hiring native speakers does not fix it

This is the question everyone asks, in a tone suggesting the answer is obvious. Three reasons it is no.

  • There are nowhere near enough of them. Germany alone employs tens of thousands of English teachers; add France, Poland, Italy, Spain, Turkey, Brazil, China. Germany could not staff the world's German lessons either. Almost every language on earth is taught, almost everywhere, mostly by people who learned it.
  • Speaking a language is not a teaching qualification. A teacher who learned it knows which mistake is coming next week. A French teacher of German who once fought the case system can see it arriving. A native speaker has often never consciously known the rule at all.
  • The big state schemes did not produce the jump. Japan's assistant-teacher programme has placed 82,628 people in Japanese schools since 1987, and the national figure has not visibly moved.

Alongside a trained local teacher a native speaker is worth having, as input the class cannot answer in the mother tongue. Instead of one, much less. The question blames the accent at the front of the room, when the problem is the number of minutes each pupil speaks. Replace the teacher and the arithmetic is unchanged.

Every system that does better bought extra hours

The Netherlands teaches other subjects in the language. In 2021, 134 of its 648 secondary schools ran an accredited Dutch-English bilingual programme. Those pupils reach higher English levels, faster, than their mainstream classmates, and the two fears people raise have not materialised: results in geography or history are not measurably different, and their Dutch does not suffer. One caution, stated by the researchers who study it: these streams attract motivated pupils from supportive families, so the raw gap flatters them.

Luxembourg is the warning. In the most multilingual school system in Europe, children from French-speaking and Portuguese-speaking homes do worse in German and Luxembourgish listening whatever their background, and worse in maths too, because maths arrives through German. In July 2025 the government put forward a law to let families choose French-language literacy instead. Teaching a subject through a language sorts children by the language they arrived with, and a system that does not plan for that turns a language policy into an inequality.

And the reform everybody adopted did the least. Starting earlier is now nearly universal, yet between 2013 and 2020 the share of lower secondary pupils learning two or more foreign languages moved by 0.8 percentage points. More years of thin contact is the one thing the arithmetic says will not work.

Three changes that would actually move the number

1. Examine the thing you want people to be able to do

Speaking has to be in the school-leaving exam, weighted heavily enough that no department can quietly skip it. It is the cheapest lever in education: it reaches every classroom in a country at once, and nobody has to be persuaded of anything.

The objections are real: speaking tests are slow, marking them consistently is hard, and a nervous pupil in front of an examiner is being tested on something other than language. But plenty of places do it, and the recipe is known: orals run inside the school, moderated from outside, recorded so a sample can be marked again. Keep it inside the school system, or leave it alone.

The knock-on effect matters more than the exam. Once speaking is examined, pair work stops being something good teachers do when they can afford it and becomes the centre of the lesson. The 43 seconds turn into ten minutes. That is the whole game.

2. Buy the hours, from other subjects and from the evenings

No minister can find three hundred extra language hours in a full timetable, and every attempt turns into a fight with maths, history and sport that the language department loses. So take the hours from a subject you are already teaching, by teaching it in that language.

The real constraint is teacher supply: a geography teacher who can teach geography in fluent English is rarer than either a geographer or an English teacher, which is why this is a stream in 134 schools rather than the design of all 648. So the answer is a national qualification rather than a national order: pay serving subject teachers to reach the level and be certified.

Then there is the free half, which costs no teacher at all. Subtitling needs a broadcasting rule rather than a budget line, and today it mostly needs defaults: whether a streaming service puts the original audio one tap away or four, and whether anyone has ever shown a fourteen-year-old how to change it. That last one a teacher can do on a Tuesday afternoon, and it is worth more than the lesson it interrupts.

3. Stop promising fluency, and hand over the tools instead

Reading a novel without constant help takes something like 8,000 to 9,000 word families, and following ordinary speech around 6,000 to 7,000. A learner at B2 carries roughly 3,250 to 3,750 of the 5,000 most frequent words. Most school leavers are well below that. School gets a pupil about a third of the way there, at the exact moment the lessons stop.

Calling that a failure is unfair to teachers who did what 45 minutes allowed. Calling it finished is dishonest to pupils. So change what the subject promises: a solid core, plus a working method for carrying on, taught as examinable content rather than a motivational poster in week one. How to keep a record of the words you personally run into. How spaced repetition works and why cramming does not survive the summer. How to find a conversation partner and keep them.

That last part is where we work, and deliberately a small part. Vokabulo exists for the several thousand words that belong to one person's life rather than to any syllabus, kept with the context you met them in and brought back on a spaced repetition schedule. It is no substitute for a teacher.

So, has it changed?

Both halves are true. The doctrine has been rebuilt four or five times since 1882, and the lesson still looks much as it always did, because what a language lesson looks like is set by 45 minutes, 28 chairs, one adult and an exam that measures reading.

The countries that do better found more hours: in other subjects, on television, in the plain fact that a Dutch teenager will need English and knows it. Where ordinary life does not give those hours away, they have to be bought on purpose. Otherwise the outcome stays what it has been for a century: a nation of people who studied a language for eight years and apologise for not speaking it.

More on the hours problem in why so many people fail with language apps, on the levels in understanding CEFR language levels, on where school leavers get stuck in breaking the intermediate plateau, and on picking a language up decades later in learning a language in later life.