Why Your English Stopped Improving and What to Do

What is the intermediate plateau, exactly?

Most learners describe it the same way: for the first year, every week felt like progress. New words stuck, grammar finally clicked, and conversations got easier. Then, somewhere between B1 and B2, the feeling of movement stopped. You can hold a conversation, watch a show with subtitles, and read an article, but you are not clearly better than you were six months ago.

This is the intermediate plateau, and it is one of the most documented phenomena in second language acquisition. Learners at the intermediate level typically need far more exposure and practice to gain the same measurable improvement that a beginner gets from a single lesson. The beginner learns "the past tense" and it changes everything they can say. The intermediate learner absorbs hundreds of new collocations and barely notices.

The plateau is not a sign that you have hit your ceiling. It is a sign that the strategy that got you here has stopped being the right strategy for where you are now. The good news is that the causes are specific, which means the fixes are too.

Why Your English Stopped Improving and What to Do

The core reason is that language learning is not a straight line of effort in, ability out. Early progress is fast because everything is new and high-frequency. The first thousand words cover most of what you hear, so every lesson feels like a leap. By the intermediate level, you are learning the second and third thousand words, which appear far less often, and you are learning to combine what you know in more precise ways.

What used to be learning has quietly become maintenance. You read the same kind of articles, watch the same kind of shows, and have the same kind of conversations, so you rehearse the language you already own instead of acquiring new language. Research on the intermediate plateau in language learning shows that words are only learned when they are encountered in meaningful contexts with enough frequency and attention. Passive, comfortable exposure does not meet that bar.

The uncomfortable truth is that feeling comfortable in English is often the exact reason you are stuck. If nothing you read or hear makes you pause, you are not at the edge of your ability, and the edge is where growth happens.

Is your input too easy now?

Stephen Krashen's input hypothesis argues that acquisition happens when you understand language slightly above your current level, what he called "i+1." The key word is slightly. Too far above and you drown; too far below and you coast. Most plateaued learners have drifted into coasting territory without realising it, because the content they chose at B1 is still the content they consume at B2.

There is a simple test. If you can follow a podcast, a series, or an article without ever stopping, rewinding, or looking something up, it is below your learning level. It is pleasant, and it keeps your English warm, but it does not grow it. A useful rule is that roughly one word or structure in every ten to fifteen should be new or uncertain.

The fix is not to abandon everything you enjoy. It is to deliberately add one source that sits just above you: a podcast aimed at native speakers on a topic you know well, a novel instead of graded readers, or news analysis instead of short headlines. The discomfort of a few unknown words per sentence is the signal that acquisition is happening.

Are you reading and listening but never speaking?

Input builds understanding, but understanding and speaking are different skills with different memory demands. When you read, the word is given to you and you recognise it. When you speak, you have to retrieve the word from memory, assemble it grammatically, and pronounce it, all in real time. Recognition and production are separate processes, and a learner who only consumes will develop a wide gap between what they can understand and what they can say.

This gap is one of the most common causes of the plateau. You feel like you are improving because you understand more, but your speaking has barely moved. The only way to close the gap is to produce language under the same time pressure you will face in real conversation.

Speaking practice does not require a native speaker to be sitting across from you. What it requires is regular, low-stakes output where you are forced to retrieve words and form sentences without minutes to prepare. Recording yourself answering a question, summarising an article you just read, or describing your day out loud all count. The key is that the words come from your memory, not from the page.

For many learners, the hardest part is simply having someone to talk to every day, which is where an AI English conversation practice partner changes the arithmetic. A conversation that can happen at any hour, without scheduling or embarrassment, removes the single biggest excuse for not speaking. But the principle stands regardless of the tool: output must be daily, and it must be retrieval, not reading aloud.

Do you get feedback on your mistakes?

Here is the uncomfortable part of the plateau: you are making mistakes, and you have been making the same ones for a long time. Without correction, an error repeated hundreds of times becomes fossilised, which means it feels natural and correct to you even though it is wrong. Fossilised errors are much harder to fix than fresh ones, which is why feedback matters more at the intermediate level, not less.

Most intermediate learners get almost no feedback. Conversation partners are polite and will not correct you, because the social cost is too high. Self-study materials grade your answers but rarely explain why something was wrong. The result is that you keep producing the same preposition errors, the same article mistakes, and the same word-order slip-ups, year after year.

What you need is feedback that is specific and repeated. It is not enough to be told "that's wrong." You need to see the correct form, understand why, and then encounter the same structure again later to check whether it stuck. Studies on retrieval practice and vocabulary learning are clear on this: testing yourself on the same material after a delay strengthens memory far more than rereading it. Correction followed by spaced re-exposure is the mechanism that actually fixes a fossilised error.

This is where the design of a tool matters. Tutoro tracks the specific words and grammar you get wrong and brings them back inside later conversations, so a weak item resurfaces on its own rather than vanishing after one correction. That is the difference between being told about an error and actually eliminating it.

What does a recovery plan look like?

The plateau is not broken by one dramatic change. It is broken by changing three things at once, each of which targets a specific cause.

First, raise the difficulty of your input. Pick one source that forces you to pause and look things up. Second, make output a daily habit, even if it is ten minutes of speaking to yourself or to a conversation partner. Third, get feedback on your specific errors and re-test them after a delay.

You do not need to overhaul your whole routine. If you currently spend an hour a day on English, spend twenty minutes on harder input, twenty on speaking, and twenty on reviewing your own mistakes. The composition of the hour matters more than its length. A learner doing those three things for an hour will outpace a learner doing comfortable input for three.

Tutoro runs this recovery loop as a single conversation. A lesson introduces vocabulary and grammar at your level, then asks you to produce it in the same thread, and the tutor reads your replies and adapts, bringing weak words back days later. It suits learners who are stuck at B1 or B2 and need structured output plus correction, and it is less useful for complete beginners or anyone who only wants to watch and listen passively.

How do you know it is working?

Progress at the intermediate level is slow enough that you cannot feel it day to day, which is exactly why so many learners give up. The fix is to measure it in a way that is sensitive to small changes.

One reliable measure is your own recordings. Once a month, record yourself speaking on the same topic for two minutes and save the file. After three months, listen back to the first recording. You will hear differences you could not feel while they were happening: fewer pauses, more varied vocabulary, better flow. This is the single most convincing proof that the plateau is moving, and it costs nothing.

Another measure is your lookup rate. Count how many words you had to look up in a week of reading and listening. If the number is zero, your input is too easy. If it is dropping over time on the same source, you are acquiring. And if you can now say something in conversation that you could only understand a month ago, that gap between recognition and production is closing.

The plateau feels permanent precisely because the changes are too small to notice in the moment. But they compound. The learner who keeps pushing slightly harder input, daily output, and corrected errors will look back in six months and see a different speaker. The learner who waits to feel motivated will still be waiting.

What if nothing seems to work?

If you have genuinely tried harder input, daily speaking, and feedback, and six months have passed with no change, the problem is usually one of two things. Either your practice is not as consistent as you believe, or you are measuring the wrong thing.

Consistency beats intensity in language learning. Twenty minutes a day for a month will move you further than three hours every Sunday, because memory consolidation happens between sessions, not during them. If your practice is sporadic, no method will save you. Fix the schedule before you change the method.

The other possibility is that you are measuring fluency by how it feels, and fluency always feels worse than it is. Record yourself, compare over time, and let the evidence decide. If the recordings show movement, the plateau is already behind you and you simply have not noticed it yet.

Should you change your tools?

Sometimes the plateau is not about you at all. It is about the tool you are using having a ceiling. Flashcard apps are superb for the first two thousand words and then become a grind with diminishing returns. Audio-only courses build listening but cannot correct your speech. Grammar textbooks explain rules but do not force you to use them under time pressure.

If your current tool only does one thing, and that one thing is the thing you are already good at, the tool is now part of the problem. The most useful question to ask of any resource is: does it make me produce language, and does it tell me what I got wrong? If the answer to both is no, it is a maintenance tool, not a growth tool.

Tutoro is built around that distinction. Lessons are authored to move you through a curriculum rather than improvising, and every exercise asks you to produce or recall something, with weak items returning until they stick. It is one answer to the plateau, not the only one, and it works best when paired with the harder input and self-recording habits described above.