Why I Never Use The CEFR Framework to Measure my Chinese Proficiency

Many learners like to describe their language level using the CEFR framework. They might say ‘I’m B1 in Mandarin’ or ‘I’m C2 in French’ and pin these statements to their Twitter bio. This is often based on what their teacher has told them or a proficiency test they’ve passed. In this blog I explain why I don’t think using the CEFR to describe our language proficiency is useful or meaningful and suggest what we should do instead: Why I Never Use The CEFR Framework to Measure my Chinese Proficiency – I'm Learning Mandarin


I haven’t read the article, but regarding the Twitter thing: I feel like it’s best to just let people say/be proud of what they like about their own achievements/perceived achievements, and not be a gatekeeper about it.


A comprehensive refutation of your article (explicit language warning):

https://youtu.be/XeLpZMuCdpU?si=LaEBdcL3oFzKkYTh (fixed link)



It would be helpful - to me for one and perhaps others too - if you could present your arguments here instead of referring us to a blog post.

If you humour me and present them here, I’d be very happy to read them. It’s entirely up to you and I quite understand if you choose not to.

I am sure that you are aware that depending upon the target language, the exams at CERF level C1 are different.

For those learning English, the CAE C1 exam is the same throughout the world - you’ll get the same C1 exam whether you sit it in Canada, Cameroon, France, Haiti, or Vietnam.

This is not the case with the Chinese C1 exam - be it Mandarin or Cantonese - you’ll still get four parts: reading listening, writing and speaking - but the exam will be totally different depending on whether you are for example, a diplomat or not.

You could very well pass the C1 exam in Chinese and function perfectly well in a legal sphere in a courtroom and yet be quite unable to order a meal in a restaurant because your C1 exam was oriented towards your future job.

Take-home message:
The C1 CAE English exam is universal worldwide, the C1 Chinese exam is not.


For example, as I wrote in the other post, in online group classes there should be a differentiator that allows participants to meet people of roughly similar levels.

Anyway, thanks for sharing your thoughts on this. If anything, it makes us rethink some “obvious” ideas that may in fact be quite arbitrary.


I didn’t say they shouldn’t put it on Twitter if they want to. I just made an argument that I don’t think it’s meaningful. I’m entitled to think that. Have a read and see if you agree.

Helpful to who? Not me!

Seriously though is it that hard to click a link? XD I mean totally fine if you don’t but you sound like you’re interested in the topic so don’t know why you’d want to punish yourself!

Is it really that hard to memorize a handful of fixed sentences and collocations for a Phd when ordering food in a restaurant or buying a train ticket?

My humble guess is it is not that hard.


Helpful to who? Not me!

Does this mean that you don’t want to have discussion here?
At first I assumed that’s why you started the thread, to encourage discussion, but the above comment suggests my inference may have been poor.
Are you instead looking to get more traffic to your blog?


I want to spark discussion. But I fail to see the issue with posting links to my blog. I asked LingQ for permission to do that first.

Looks good. I sent your blog post and contact info to the Council of Europe to let them know the CEFR is meaningless for Chinese.



Toby, you could be more specific here:
It’s the “Chinese Council of Europe” :slight_smile:

1 Like

Hi Michilini,

I agree, the CEFR framework is problematic - for some of the reasons you mention in your blog post.
However, you seem to overgeneralize your UK experience in Chinese:

  1. My experience with Romance languages in Germany, France and Spain is completely different:
    Students who study foreign languages at a German university are usually expected to study abroad (i.e., in the countries where their target languages are spoken) for 1-2 semester(s).

So, if they are able to give an academic presentation, they usually have also no or at least not many problems ordering a meal in a restaurant in their L2 - esp. after studying abroad for 6-12 months.

That may be different in the UK reg. Mandarin, but that’s a negligible minority of all students in the UK:

  • According to Gov.UK “Education and training statistics for the UK” (for 2022), there were ca. 2.9 million students in higher education in the UK: Education and training statistics for the UK, Reporting year 2022 – Explore education statistics – GOV.UK
  • Let’s say, there were ca. 5000 people studying Mandarin at British universities in 2022 (this number is probably too high, see the “Report on the Present State of China-related Studies in the UK” (10/2019)).
  • Let’s say 10 percent are doing a PhD in Chinese studies and 50 percent of those PhD students have never been to China, so they might have trouble ordering a meal, then we’re talking about 250 people here.
    You may have even met a handful of those PhD students who have trouble ordering a meal in person… But more than 10? :slight_smile:

“For example, learners who meet C2 criteria and can deliver academic presentations in a foreign language often struggle to meet the A2 criterion above when ordering food at a Chinese restaurant or buying a train ticket.” (Michilini - highlighting by me).

Therefore, your paragraph translates to:
The tiny minority of learners of Chinese in the UK who meets C2 criteria (esp. the 250 PhD students who have never been to China) and can deliver academic presentations in Chinese sometimes struggles to meet the A2 criterion above…

  1. But can you talk about “every” imaginable topic in your L2 in a competent way?
    Of course not - that’s not even possible for native speakers -, because the important thing here is that you know your stuff.

In other words, fluency without competence in a subject matter XYZ is just incompetence in XYZ (and that’s true both for native and non-native speakers).
The only unfair advantage that native speakers, esp. academics, have in this context is that they have more words to hide their incompetence. Ergo: It’s “fluent incompetence” :-o

  1. “they would, on what basis do they decide which topics a C2 learner should be able to understand and which topics they can get away with not understanding? If they wouldn’t, who on earth does qualify as C2?” (Michilini)

Insofar as there are countless (specialized) topics in our hyper-complex society, even the best of the best of our species can only appropriate a few of them (with the exception of Dunning-Kruger sages like Donald Trump: they can even talk about the “unknown unknowns” with confidence and ease. That’s definitely C triple plus).

So you always have to prepare your topics…, even as a native speaker.
The important point here is the presentation, which you usually have to prepare as well and which should consist of similar verbal and non-verbal nuances as educated native speakers (would) use.

See the qualitative aspects of oral language use: Qualitative aspects of spoken language use - Table 3 (CEFR 3.3): Common Reference levels - Common European Framework of Reference for Languages (CEFR)

  1. BTW, my main criticism of the CEFR framework is that they don’t use commas /dots and digits after the comma / dot. I mean how do you express that your actual language level is “A1,45 or C2.39”?

But that’s typical of the humanities, isn’t it? :-0

PS -
My German colleague at work did his PhD in the UK.
He speaks Japanese, but struggles both with ordering meals and delivering
academic performances in Japanese.

Ergo, the anecdotal evidence detector tells us that the UK is not a good place for learning Asian languages. What’s the reason for that: the weather, the food, the end of the British Empire, Boris Johnson and his parties, Manchester City, …?

Ah, I forgot to mention that he did his PhD in computer science using the English language.

Yeah, learning Japanese in the UK just sucks, baby. That should be mentioned in the Japanese part of the CEFR framework as well, don’t you think?


D1 is the level to go. This is the level where even the most educated native speakers begin to understand that they still suck at their L1.

And D2 is reserved for the AI superintelligence to come…

  • speaking
  • listening
  • reading
  • writing
  • nonverbal skills (gestures, etc.)
  • grammar
  • pronunciation
  • knowledge of the style of communication (Dutch / German / Israeli, etc. directness vs Asian indirectness, the styles of humor, politeness strategies, etc.)
  • cultural background knowledge
  • your topic-specific expertise (esp. if it’s job related).
  • months of living in the foreign country where the L2 is spoken
    You could also prove that you’ve studied abroad, as it is assumed that you need an advanced level in your L2 to study at a university (and yes, Michilini, that includes being able to order a meal unless you want to starve in your host country :-)).

Apart from that ,it may be better to adopt a SW perspective on “testing”:
L2 tests don’t show our exact level of proficiency, but they can show us some of our weaknesses in certain SLA (multi) dimensions.

It’s similar with SW tests: tests show us the presence of some, but not the absence of all errors. See, for instance: software - Testing can detect the presence of error but not the absence of error, why? - Software Engineering Stack Exchange


@Michilini I’m glad this wasn’t like the usual clickbait change-my-mind kind of stuff you so like.

You bring up some good points, but at the end of the day the CEFR works reasonably alright at its job. Just as there are downsides to using the CEFR, there are also downsides in using your personal qualitative description. Namely, (1) it’s longer, (2) it’s not universal/easy to compare people.

When someone tells me they are C1, I generally understand what they mean. Obviously, there is some variance that maybe they don’t know food vocabulary or maybe their listening skills are great but they are poor readers, but this is to be expected. What I do know for sure though is that they are quite good in the language. If I’m interested, like as a follow-up to your personal qualitative description, I can also ask for further details, such as “Do you read books in your L2?”, “Do you watch movies in your L2? With or without subtitles?”, “Have you ever lived in/visited the country, where they speak the language natively?” Etc.


I agree that ultimately we react to the “vagueness of proficiency” with the “vagueness of evaluation” (three categories, i.e., beginner - intermediate - advanced, which have to be taken “cum grano salis”).

If you want to know details, you need to dive deep into the multidimensions mentioned below (see my comment to Sergey). In many everyday communication situations, that’s overkill, though.

It’s the same in many other practical skills acquisition processes (math, calisthenics, etc), btw, …

@Michilini I’m glad this wasn’t like the usual clickbait change-my-mind kind of stuff you so like.”
You mean his “troll rating” for this post is not C1.7, but just A0.9? :wink:

It’s the algorithm, not Michilini that is fond of clickbaiting titles :slight_smile:
However. Michilini’s problem seems to be more his penchant for

  • “common sense” (failing to understand that the line between “sense” and “nonsense” is very thin in this area, which is more than surprising for someone with a background in philosophy),
  • “anecdotal evidence”,
  • “overgeneralizations”,
  • the construction of (imaginary) “SLA enemies” such as Krashen drones aka input fanatics,
  • the construction of pseudo-problems (e.g. the degree of self-worth / respect related to LingQ stats)

More scientific reading (especially about SLA research) might help, alas… Well, hope dies last :slight_smile:

A nice WE to you all,
Peter out


Idk, I think half my close friends have PhDs (including my brother) and I’m increasingly convinced that they shouldn’t be allowed to talk about anything they didn’t get a PhD in. Unless they got a PhD in ordering coffee (in their TL), I am certain they might not be able to do it.

That’s just how it goes, forget what it’s like to be a normie, and now you can’t order coffee.

Edit: one of these PhDs actually lives in Germany, and definitely cannot order coffee, nor a Döner Kebab, auf Deutsch. I did convince him to use Lingq for a week, he still cannot order his Kebab.


I realized I hit a D1 in Norwegian when someone called me a “mjølk ass” in Norwegian after I told him he was wasting his time on Duolingo. My wife was like, why do you keep talking to this guy, and I’m like hang on, I’m passing my D1 Norsk eksam.