At C1 you can express any idea you have. What is missing is not vocabulary or grammar but knowing which of the many correct options a proficient speaker would actually pick.
C1 is where most serious learners stop. Not because they give up, but because by then everything works: you read books, watch films without subtitles, argue a position, do your job in the language. And then, year after year, very little changes.
The stretch from C1 to C2 differs in kind from every stretch before it. Understanding how it differs is the precondition for crossing it.
This is the most widespread misunderstanding about the framework, and the Council of Europe says the opposite explicitly. In the CEFR Companion Volume they write that level C2, although it is termed Mastery, is not intended to imply native-speaker or near native-speaker competence. What it describes instead is the degree of precision, appropriateness and ease that characterises the speech of highly successful learners.
Those three words repay slow reading: precision, appropriateness, ease. None of them is native-like.
This matters practically, not merely as consolation. If your target is a native speaker, you are chasing something unmeasurable and unreachable, and the chase is demoralising. If your target is precision, appropriateness and ease, those are three trainable things — and they are trained in three different ways.
In 1983 the linguists Andrew Pawley and Frances Syder posed a problem that still describes the C1-to-C2 stretch better than anything since. They called it nativelike selection.
For almost any idea you want to express, there are dozens of grammatically correct ways to say it. Proficient speakers use one or two of them. The rest are not wrong — they are simply the ways nobody says it.
A C1 learner has enough grammar and enough vocabulary to construct all the dozens. What they lack is knowing which one is used.
In English, both of these are grammatically flawless: I want to marry you and Will you marry me? Only one is what people actually say when proposing. No rule of grammar rules the other out. You either know it or you do not.
This is why the stretch cannot be solved by studying harder in the way you are used to studying. It is not a knowledge gap. It is a gap in how many times you have met the right version.
One: collocation. The words are right individually and wrong together. English says heavy rain, never strong rain, though both would mean a lot of rain. No logic accounts for it. A C1 learner knows heavy, knows strong, knows rain, and still assembles it wrongly, because they learned the words separately and combined them themselves.
Two: register. At C1 you know what to say. At C2 you know where it belongs. Every language splits between speech and writing, intimate and formal, and the right word in the wrong register is still the wrong word. This is the hardest layer to learn from books, because books rarely mark the boundary.
Three: automaticity. At C1 you can produce the correct sentence — but it costs effort, and a listener can feel the effort. This is precisely the ease in the Council of Europe's definition. In skill-acquisition terms it is the move from knowledge you have to an operation you perform without thinking, and it comes from deliberate repetition rather than further understanding.
There is a mathematical reason effort stops paying at exactly this point, and knowing it saves you from blaming yourself.
To read a text without assistance you need to recognise roughly 98 per cent of the words in it — the figure comes from Hu and Nation's research. Reaching that threshold for general written text is usually estimated at around eight to nine thousand word families. A C1 learner is typically close to that. C2 needs considerably more.
The problem is the nature of the words that remain. By definition they are rare. At two thousand words you meet something new every few sentences. At ten thousand you can read a whole book without meeting again the particular word you were trying to fix.
That breaks the very mechanism that carried you here. Spaced repetition needs repetitions; extensive reading needs density. At C1 the density disappears.
This is the real reason the climb takes years. Not that you have become a worse learner. Each remaining word simply shows up far less often.
There is a second reason, and it sits somewhere unexpected: success itself removes the pressure to improve.
Fossilisation describes errors that have settled and no longer disappear on their own. At lower levels errors cause misunderstanding, so others correct you and you correct yourself. At C1 your errors no longer obstruct communication. Listeners understand what you meant and move on. Nobody corrects someone who speaks their language well — it would sound like nitpicking.
Set that beside the noticing hypothesis: you do not acquire what you do not notice. At C1 nothing forces you to notice any more. You produce the wrong collocation, you are understood, the conversation passes, and the error is reinforced once more.
Which means that at this stage you have to manufacture the pressure that circumstances used to supply for free.
One: narrow reading, not extensive reading. Here the advice reverses. Extensive reading — a lot of easy, varied material — is what carried you from B1 to C1, and there it was right. At C1, do the opposite: read narrowly. Same author, same subject, same field, for weeks at a stretch.
The reason is the frequency problem above. Rare vocabulary is not spread evenly — it clusters by domain. Ten articles on architecture will show you the same rare words often enough for them to stick; ten articles on ten subjects will not. Narrow reading manufactures the density that C1 has lost.
Two: keep a collocation notebook, not a vocabulary list. At this level recording a single word is nearly useless — you already know its meaning. What you do not know is what it goes with. Record the chunk, and record the sentence you met it in.
Three: back-translation. This is the exercise that exposes the selection gap most sharply. Take a passage written by a proficient speaker and translate it into your own language. Leave it for a few days. Then translate it back without looking at the original. Now compare.
Your version will be grammatical. It will also be different. Those differences are your syllabus — they mark precisely the places where you chose a correct option that was not the used one.
Four: write, and be corrected. Because self-correction has stopped working, you need an external source of it. One thoroughly corrected page is worth twenty unread ones. And ask for the reason rather than accepting the fix: the learning is in the answer because we don't say it that way.
Five: record yourself. Speak for a minute or two, then listen back. It is deeply uncomfortable and remarkably effective. You will hear where you hesitate, where you reach for the same structure repeatedly, where your intonation runs the wrong way — none of which you can perceive while speaking.
Six: pick a domain. C2 across all subjects is not a real thing. Even native speakers are only genuinely proficient within their own fields — a doctor and a lawyer cannot read each other's documents fluently. Choose two or three areas you actually care about and go deep there. Deep in a few places is achievable; broad everywhere is not.
B2 to C1 can take one to two years of steady work with real immersion in the language.
C1 to C2 commonly takes three to five years or more, and a great many people never arrive. That figure is not there to discourage you. It is there so you do not conclude something is wrong with you when six months pass and nothing seems to have changed. At this stage six months is simply too short an interval to see a difference.
Most practical purposes are already met at C1. University study, professional work, reading in your field, living abroad, making friends — C1 does all of it.
C2 is worth pursuing if you teach the language, translate, write professionally in it, or work somewhere a wrong nuance carries a real cost. Outside those cases, the time it takes to go from C1 to C2 in one language usually returns more if spent elsewhere — including on taking a second language to B2.
Check what you actually need first. It is usually lower than you assume.
Not that you sound native. Plenty of C2 speakers have a pronounced accent, and it has nothing to do with the level.
What changes is that the language stops being a subject you study. You stop translating, stop selecting, stop monitoring yourself as you speak. It becomes something you think inside, rather than something you use to restate what you already thought in your first language.
That is the ease the Council of Europe put into the definition. It turns out to be the hardest of the three words to earn.
The same article, told in plain words — for younger readers, or for anyone who wants the point quickly.
C1 is where most serious learners stop. Not because they give up, but because by then everything works: you read books, watch films without subtitles, argue a position, do your job in the language. And then, year after year, very little changes.
This is the most widespread misunderstanding, and the framework says the opposite explicitly. C2 is not intended to imply native-speaker competence. What it describes is precision, appropriateness and ease.
Read those three words slowly. None of them is native-like.
This matters practically, not merely as consolation. If your target is a native speaker, you are chasing something unmeasurable and unreachable. If your target is those three things, they are trainable — and trained in three different ways.
For almost any idea, there are dozens of grammatically correct ways to say it. Proficient speakers use one or two. The rest are not wrong — they are simply the ways nobody says it.
A C1 learner has enough grammar to construct all the dozens. What they lack is knowing which one is used.
In English, both of these are flawless: I want to marry you and Will you marry me? Only one is what people actually say when proposing. No rule of grammar rules the other out. You either know it or you do not.
Which is why this stretch cannot be solved by studying harder in the way you are used to. It is not a knowledge gap. It is a gap in how many times you have met the right version.
The words that remain are, by definition, rare. At two thousand words you meet something new every few sentences. At ten thousand you can read a whole book without meeting again the particular word you were trying to fix.
That breaks the very mechanism that carried you here. Spaced repetition needs repetitions; wide reading needs density. At C1 the density disappears.
Not that you have become a worse learner. Each remaining word simply shows up far less often.
There is a second reason, and it sits somewhere unexpected: success itself removes the pressure to improve.
At lower levels errors cause misunderstanding, so others correct you. At C1 your errors no longer obstruct anything. Listeners understand what you meant and move on. Nobody corrects someone who speaks their language well — it would sound like nitpicking.
So at this stage you have to manufacture the pressure that circumstances used to supply for free.
1. Narrow reading, not wide reading. Here the advice reverses. Wide, easy, varied reading carried you from B1 to C1, and there it was right. At C1, do the opposite: same author, same subject, same field, for weeks at a stretch. Rare vocabulary is not spread evenly — it clusters by domain. Narrow reading manufactures the density that C1 has lost.
2. Keep a collocation notebook, not a vocabulary list. Recording a single word is nearly useless now; you already know its meaning. What you do not know is what it goes with.
3. Back-translation. Take a passage by a proficient writer, translate it into your own language, leave it a few days, then translate it back without looking. Now compare. Your version will be grammatical — and every place it differs is a place where you selected and they knew.
4. Write, and be corrected. Self-correction has stopped working, so you need it from outside. One thoroughly corrected page is worth twenty unread ones. And ask why, rather than accepting the fix — the learning is in the answer because we don't say it that way.
5. Record yourself. Deeply uncomfortable and remarkably effective. You will hear where you hesitate, where you reach for the same structure again, where your intonation runs the wrong way — none of which you can perceive while speaking.
6. Pick a domain. C2 across all subjects is not a real thing. Even native speakers are only genuinely proficient inside their own fields — a doctor and a lawyer cannot read each other's documents fluently. Deep in a few places is achievable; broad everywhere is not.
B2 to C1: one to two years of steady work. C1 to C2: commonly three to five years or more, and a great many people never arrive.
That figure is not there to discourage you. It is there so you do not conclude something is wrong with you when six months pass and nothing seems to have changed. At this stage six months is simply too short an interval to see a difference.
Most practical purposes are already met at C1 — university, professional work, reading in your field, living abroad, making friends.
C2 is worth pursuing if you teach the language, translate, write professionally in it, or work somewhere a wrong nuance carries a real cost. Outside those cases, the time it takes usually returns more if spent elsewhere — including on taking a second language to B2.
Not that you sound native. Plenty of C2 speakers have a pronounced accent, and it has nothing to do with the level.
What changes is that the language stops being a subject you study. You stop translating, stop selecting, stop monitoring yourself as you speak. It becomes something you think inside, rather than something you use to restate what you already thought in your first language.
That is the ease in the definition. It turns out to be the hardest of the three words to earn.
These articles summarize well-established research in learning science and linguistics. Key sources and further reading:
The question assumes focus is a muscle you train. The strongest evidence says the training barely carries over, and the...
No time, no talent, too old — those are explanations, not causes. What can be measured is duller: one sentence with a ti...
Not one event but four, running on clocks a billion-fold apart: a wave at 400 milliseconds, a word locked in overnight,...