I interviewed a guy who learned to speak utterly phenomenal Chinese within 1 year while living in the UK. Here's how he did it

Open Forum
41

Right. Also let’s be careful not to make the assumption that Will’s speaking skills translate to full listening skills beyond conversation. Can he understand Chinese soap operas? Can he understand Chinese humor? Can he understand Chinese TV Shows? Can he understand Chinese movies?
Also: what is his level of reading and writing?

Those are important data points too.

If we’re going to make a true apples to apples comparison we need to control for that.

I bet many of the Mandarin learners who don’t make the same speaking progress as Will have tried to learn the writing system. How does Will compare to them? As in, is his understanding of the writing system and his reading comprehension comparable? I bet not.

This is not to disparage Will because it is broadly the same principle that I use: focus on the spoken material only in order to avoid the pitfalls of trying to learn a new writing system and reading at the same time.

What I’d therefore like to point out is that I’m barely literate in my own target distant language (Russian) as a result of my focus on listening/consuming video input.

What makes Will sound so impressive is he appears not only to be fluent but that his pronunciation and grasp of accurate grammar seems much more advanced than others who have put in similar effort.

There are a couple statements Will makes that are inaccurate however. He says he learned it will still being in England and that he didn’t focus on grammar at all.

Although he was in England, he was in a Mandarin speaking environment in which he was actively engaged. Therefore it’s not true to draw the conclusion that he was essentially still in England.

Although he didn’t take formal grammar classes he did in fact concentrate on grammar by the use of sentences as models. This is the basis of the glossika method also.

42

Is he though? Does he differ from “all others”?
Luca Lampariello. Idahosa Ness. Gabriel Wyner. They all have awesome speaking skills.

That does not mean that the data points you give are inaccurate, just that you are drawing different conclusions even though you are stating accurate data points.

I think you’re almost in agreement though. Of the above statements I think the only thing to quibble with is “regardless of time spent on task”.

It is clear to me at least that if you try to learn to read in a distant language while paying no attention to listening you will have close to zero listening comprehension. You can find evidence of that on these forums. People with really high numbers of words read but who cannot understand the spoken language. In my case it is the opposite, I can understand the spoken form but can barely read.

I think you’re correct in the why, however: his specific combination is precisely tuned to his target of conversational fluency.

But… I don’t think it is accurate to say that he is much more fluent in all areas regardless of effort if you are in fact saying that.

43

“My grammar also sucks”
“Grammar light” approaches are really helpful for understanding distant L2s at the beginning. For example,

  • Michel Thomas
  • Language Transfer (https://www.languagetransfer.org/)
  • short grammars in phrasebooks, on LingQ, etc.
  • Memrise decks for conjugation or grammar pattern drills
    etc.

For non-Indo-European languages to be acquired, I would never start my language learning journey without such approaches. However, for related L2 languages, they can often be ignored.

44

“IPA, early speaking and IMMERSION combined with anki sentences and audio/video input.”
If we want to generalize here, my hypothesis is that the concrete mix depends on three dimensions:

  1. The distance of the L2 (the familiarity of the vocabulary, etc.)
  2. The complexity of the pronunciation (tones, pitch accents, click sounds, etc.)
  3. the complexity and distance of the writing system(s)

→ my tentative distinctions would look like this:

  • If 1), 2) and 3) are not problematic, an “(ultra)reading while listening approach” enhanced by (early) speaking/writing activities plus SRS (especially for specific collocation, idiomatic and sentence patterns) could be the way to go in many cases.
  • If 1) and 3) are not problematic, but 2) is (e.g., in the case of Indo-European speakers trying to learn the sophisticated pronunciation of European or Brazilian Portuguese), then IPA or equivalent solutions are helpful. The rest as before.
  • Especially if 3) is the main problem in combination with 1), 2) or 1) and 2), then an oral-first approach à la xxdb / Will (pronunciation practices, early deliberate speaking practices, an SRS-sentence pool relevant to everyday conversations, AV input activities, and a grammar light approach) might be the best solution for the start of our language learning journey.
45

Great, that you mention listening comprehension, this happens to be a pet peeve of mine…

Let’s make two assumptions:
1 Will spent between 1330h and 2000h actively engaged with the language (learning, conversing etc.)
2 his listening is on par with his speaking abilities, i.e. excellent

Then, I must say this number sounds extremely low to me. I’m sitting here at about 1k hours of active listening and my listening comprehension is pretty mediocre. An example from real life: last week, Apple podcasts recommended me a program called “无聊斋”, I listened to one episode 1h 30m, and I assure you, I didn’t understand a thing :slight_smile:
(Of course I’m ready to admit that I’m a particularly slow learner, who has an idiosyncratic / inefficient approach to learning languages - no doubt).
My current target is 2k hours of listening, I expect to be able to understand that podcast by then. And even then there is still a long road ahead, after all the aforementioned Khatzumoto from AJATT famously prescribed his followers 10.000 hours of listening! (this may or may not have been a facetious remark.)
Maybe this helps to put Will’s achievement into perspective, even if it comes from a sub-par learner.

46

I like it.
I think for me, I have achieved decent-ish comprehension of Russian but I don’t have the ability to output grammatically accurately and I’ve been trying to figure out how to do that without taking a formal grammar class.

It’s clear that being immersed and forced to output in order to survive works.

Without immersion though?

I think for me, I have achieved decent-ish comprehension of Russian but I don’t have the ability to output grammatically accurately and I’ve been trying to figure out how to do that without taking a formal grammar class.

What I have observed is that my brain seems to just be able to piece together the grammar passively from watching video plus occasional light reading about grammar with no attempt to memorize it.

The challenge however, specifically with my own method is that I don’t have active output models of the grammar so although my pronunciation is decent, in outputting phrases I’m stringing together words in combination with phrases I’ve heard from youtubers.

So to Will:
His pronunciation is good from his audio based approach plus IPA.
His grammar?
How to tease that out?
Is it his sentences in anki?
Is it his exposure to feedback with natives in conversation?
Is it that Mandarin grammar is not overly complex compared to say, Russian?

I don’t know the answers to these.
I’m hoping that it is the sentences in anki.

For me I know that using model sentences works after doing my method. I did this for Spanish. But unfortunately I also did almost the same thing as Will. After memorizing my initial 7000 words I did some audio courses and watched a bunch of telenovelas. Will immersed right at the beginning after the IPA whereas I immersed after I completed my initial lift. But the immersion was there. I went from simplish Spanish to near perfect Spanish in about a year. Meaning my Spanish feels similar to my English.

I can’t tease out the immersion factor. I hope it’s just the sentences because it would be hard to replicate the immersion (and to be frank I don’t have the desire to immerse). But possibly you can get part of the immersion if that’s the missing factor by doing what Steve Kaufmann does: italki etc and talking to native speakers.

47

@bamboozled:

Sorry I can’t reply directly.

“Great, that you mention listening comprehension, this happens to be a pet peeve of mine…
Let’s make two assumptions:
1 Will spent between 1330h and 2000h actively engaged with the language (learning, conversing etc.)
2 his listening is on par with his speaking abilities, i.e. excellent
Then, I must say this number sounds extremely low to me.
I’m sitting here at about 1k hours of active listening and my listening comprehension is pretty mediocre.”
Have you been using mostly lingQ reading though?
If so, it’s not an apples to apples comparison.

I have about 600-700 hours of listening to Russian and my listening comprehension is decent. I can for example understand this perfectly without subtitles:

I can’t understand netflix shows completely, although I can get the jist.

“(Of course I’m ready to admit that I’m a particularly slow learner, who has an idiosyncratic / inefficient approach to learning languages - no doubt).”
I doubt you are a slow learner. The guy’s focus is specifically on conversational fluency. Which is related to mine: my focus is specifically on understanding youtubers and netflix shows.

“My current target is 2k hours of listening, I expect to be able to understand that podcast by then.”
Can I make a suggestion?
I have used the same approach twice in a row now:
First I watch TPRS youtubers in the target language (TL).
Next I watch slow spoken youtubers in the TL.
Next I watch a mixture of slow spoken youtubers and language teacher blogs on youtube in the TL.
Next I watch regular youtubers and language teacher blogs in the TL.
Next I watch easy TV shows such as telenovelas/soap operas.
Next I watch netflix shows.

This ^^^ is in order of difficulty. My opinion/hypothesis is you can’t skip steps.
Therefore try at the bottom first?
Can you understand all the TPRS? If not, you can’t progress till you can.
Can you understand all the slow spoken youtubers? If not, you can’t progress till you can
etc etc

If what you are doing is reading with a tiny bit of listening then it just won’t work if you are trying to jump into something advanced. Listening and reading are two different skills.
NOTE: I don’t know what your method is, I’m making a wild ass guess here.

“Maybe this helps to put Will’s achievement into perspective, even if it comes from a sub-par learner.”
Nobody is mocking it. We’re taking it apart and putting it back together again. I’m very glad to see the success he has made especially since his method overlaps to some degree with my own experience.

48

@bamboozled
“I must say this number sounds extremely low to me.”
I have no idea how good Will’s listening comprehension really is.

As a non-native speaker of English, for example, it’s one thing to watch and understand “Game of Thrones” on TV. But listening to GRRM’s audiobooks is another thing entirely.
And if GRRM is too easy, then just switch to Thomas Pynchon, for example :slight_smile:
The same is true in other L2s.

So if Will is conversationally fluent (let’s say he’s a solid B2 or B2-C1), that doesn’t necessarily mean he can also understand most audiobooks (at least for a general audience) in Chinese.

And, as @xxdb mentioned, he might still have trouble understanding Chinese comedy/humor on TV, because sometimes vocabulary isn’t enough in this case, but you also need more or less profound cultural background knowledge.

49

“Also: what is his level of reading and writing?” (@xxdb)
Will said in his recent interview with Rita that he knew about 100 Chinese characters. In other words, he’s more or less illiterate. But I expected that from a strict oral-first approach.

50

I’m sorry, Michilini, but without concrete stats from Will and other advanced Mandarin learners, this is a “ghost discussion” based too much on feelings and vague notions (“fluency”, “lightning speed”, “outliers”, etc.).

“I think it has a lot to do with the combination of efficient methods and techniques he used.”
So far I haven’t seen anything “special” here: neither in the individual techniques nor in the mix. In my opinion, the recipe for his success therefore essentially lies in his strict fluency-first perspective (ignoring all activities related to reading / writing and ignoring all content that isn’t useful in conversations).

51

I think Michilini’s main gist is that it is very unusual for a mandarin learner to be able to speak this well. @michilini: Is that a fair characterization?

That said, I do agree with you that if you want to be good at speaking then not focusing on speaking won’t work. And yes, the method does overlap substantially with other methods even if it’s not a complete perfect match: (Luca Lampariello, Mimic Method, FluentForever, Benny Lewis, Glossika etc etc).

Great discussion as always though.

52

@xxdb
“Is it his sentences in anki?
Is it his exposure to feedback with natives in conversation?
Is it that Mandarin grammar is not overly complex compared to say, Russian?”
After listening to Will’s interview with Rita my guess is that it’s the combination of those three aspects.

53

“…but without concrete stats from Will and other advanced Mandarin learners, this is a “ghost discussion” based too much on feelings and vague notions…”

Well I don’t think it’s meaningful to talk about someone’s language level in any other way. Language profficiency is difficult to define but easy to identify. It’s QUALITATIVE not QUANTITATIVE.

I just don’t share your view that that makes it less real.

Anyway, believe me Will reveals loads more stuff in the actual podcast. What I’ve said today is just the tip of the iceberg!

I’ll try and get it edited & published this weekend. Stay tuned.

54

I specifically said Mandarin. Yes he speaks Mandarin more fluently than those people.

I think that what Will has done is obviously remarkable.

Yet it seems the first instinct of some people on here (I don’t mean you) is to immediately stick their fingers in their ears and shout NOTHING TO SEE HERE!

My first instinct when I see someone whose achievements outshine my own is to wonder what he’s doing differently and what I can learn from it.

That’s one of my favourite things about hosting my podcast.

Many, many people I know who’ve studied Mandarin for years still can’t string a sentence together. Expectations among native speakers are extremely low because they’re used to hearing foreigners speak Mandarin so badly they can’t even understand them.

It is a tragic state of affairs. And I think Will’s story can help a lot of people to rethink their language learning approach and realise learning to speak Mandarin to a high level is possible and needn’t take decades!

55

“I think Michilini’s main gist is that it is very unusual for a mandarin learner to be able to speak this well.”
Hard to say, but there are “famous” youtubers such as OrientalPearl (Oriental Pearl - YouTube ) or Serpentza. who have lived in China for many years, so Will can’t be one of the “happy few”.

"Language profficiency is difficult to define but easy to identify. (@Michilini)
That’s were you’re mistaken. Language experts (teachers, linguists, etc.) usually use many oral/written tests to determine

  • the size of the learners’ vocabulary,
  • the complexity of their syntax,
  • their knowledge of different language registers,
  • their sophisticated use of typical collocations,
    etc.
    And all these aspects ultimately define “language proficiency”.
    In short: Relying only on one’s “feeling for the language” isn’t good enough.

“It’s QUALITATIVE not QUANTITATIVE.”
“Statistical learning in language acquisition” (Statistical learning in language acquisition - Wikipedia) and AI tools à la “Deepl”. for example, say “You don’t know me, do you?” in this context :slight_smile:

56

@Michilini: I specifically said Mandarin. Yes he speaks Mandarin more fluently than those people.

It’s impossible for me to tell but putting the names of them into youtube with mandarin e.g. luca lampariello mandarin idahossa ness mandarin etc gives what sounds to me fluent responses. Are they as good as Will? I can’t tell. I just watched his interview with Rita and if I closed my eyes he does in fact sound Chinese to me. He even has the mannerisms.

I think that what Will has done is obviously remarkable.
Yes it is, but join the dots to your final statement: It’s possible to do.
So you need to reconcile the idea of it being “remarkable” with “it’s possible to do” IMO

Yet it seems the first instinct of some people on here (I don’t mean you) is to immediately stick their fingers in their ears and shout NOTHING TO SEE HERE!
I think you’re reading into it. Peter is saying it’s not remarkable. I’m not. I’m saying it’s epic. But I also think that Peter is right when he said that Will’s method makes sense that it would work and it’s not in fact so esoteric as it first looks. The Method that is. That everyone else is not using Will’s method or something similar is kind of esoteric. But then there are also hundreds of thousands of people who have had years of Spanish or French traditionally and learned nothing. To those people, hearing someone speak Spanish fluently who hasn’t even taken a single Spanish class would sound literally amazing to them. In that sense because Will is using the correct method and everyone else is doing it wrong is much less esoteric.

My first instinct when I see someone whose achievements outshine my own is to wonder what he’s doing differently and what I can learn from it.
Agreed. I want to know what Will’s magic sauce is. Especially since his method overlaps somewhat with my own. He has gotten to the full shebang, listening AND speaking. I have only gotten to listening in my (almost-as-difficult) target language but my speaking is nowhere near that level. I do have equivalent level to Will in Spanish, however, so I can equate my successes to his and derive a meaningful comparison.

That’s one of my favourite things about hosting my podcast.
Many, many people I know who’ve studied Mandarin for years still can’t string a sentence together. Expectations among native speakers are extremely low because they’re used to hearing foreigners speak Mandarin so badly they can’t even understand them.
The same thing is true of other distant-from-English languages. In that sense Mandarin isn’t special. As an example: even Steve Kaufmann is struggling with Arabic somewhat. And Steve knows how to do it.
But yeah. I appreciate your podcast because Mandarin is likely next on my list. I’m actually doing it a teensy bit right now as I build up my word list. So it would be good to not make mistakes in advance. The IPA thing I think is a very good catch. For example I have a seriously hard time understanding how the Mandarin ‘r’ as in ‘ren’ is actually spoken. I cannot produce it at all.

It is a tragic state of affairs. And I think Will’s story can help a lot of people to rethink their language learning approach and realise learning to speak Mandarin to a high level is possible and needn’t take decades!
I think both Peter and I as well as others agree with you on this. Language learning shouldn’t take years. And in fact it doesn’t.
FSI says that near languages to English take about 500 hours of practice all the up to distant languages which take up to a couple thousand hours.
My experience so far concurs with that:
French I did 500 hours and I understand it pretty well. I didn’t speak much so my Speaking isn’t epic but it’s OK. I didn’t have to tool up much on grammar because French grammar is a simplified analogue of Spanish grammar mostly.
Russian I have some 600 hours in and I understand it decently though not as well as French. Again I didn’t speak much and I have been severely handicapped by that and I need to overcome that so it’s interesting to me to listen to this guy’s success because Mandarin is in the same league of difficulty (perhaps a bit more) than Russian overall.

I suspect, though, that spoken Mandarin is in fact easier than spoken Russian once you get over the pronunciation hurdles because as far as I can see the grammar isn’t as complicated. I hope this is true to be honest. Will himself says in his interview with Rita that he’s scared of Spanish grammar because Chinese didn’t have that much. If he’s correct then using much of his method should work for anyone. He also did point out specifically that doing 5 minutes a day won’t work, you need to put in the hours. And that might in fact be the guts of why most folks fail: they don’t put in the work. Both for French and Russian I started out doing a minimum of 2-3 hours a day sometimes more, for the entire six month period. I’ve slacked off in Russian, I do at most an hour a day now.

Anyhow, again, thank you for posting this and doing what you do.

57

Hard to say, but there are “famous” youtubers like OrientalPearl (https://www.youtube.com/c/OrientalPearl ) or Serpentza. who have lived in China for many years, so Will can’t be one of the “happy few”.

I guess it is hard to say. Unless…you’re proficient in Mandarin in which case it isn’t hard to say at all.

Will is head and shoulders above almost all YouTubers out there in terms of his spoken fluency. Literally just go and ask any Chinese person to compare Will speaking Chinese with other popular YouTubers including the ones you mention.

It is not necessary to have any statistical knowledge in order to observe that very obvious fact.

If it wasn’t the case I wouldn’t have sought him out to interview.

58

That’s were you’re mistaken. Language experts (teachers, linguists, etc.) usually use many oral/written tests to determine

  • the size of the learners’ vocabulary,
  • the complexity of their syntax,
  • their knowledge of different language registers,
  • their sophisticated use of typical collocations,

What use is this as an argument? The fact that standardised testing exists doesn’t prove its validity or usefulness in assessing language proficiency. Nice try.

59

So far I haven’t seen nothing “special” here:

Anway don’t forget I’m the only one who knows the full facts. I’m the one who conducted the interview, remember?

I guarantee when this interview hits the airwaves the world of language learning will never be the same again…

Wait. And. See.

60

Similar to grammar, I use IPA surgically. (Also similar to grammar, the first place I go is Wikipedia.)

I might once in awhile read the Wikipedia article on phonology:

If a reference is made to a certain letter or sound then I may consult this article or the article for the letter.

To work on pronunciation, I will then record myself every week trying to say words or phrases and listen to a native speaker trying to get it as close as possible. When I notice gaps, that is when I consult IPA.

I use a site called forvo for most common words and phrases, and then whatever source I want for longer sentences.

Not just relying on myself, I have tutors and friends that are not shy to correct my pronunciation of things. In many of those cases I may not hear the difference, or worse, I am importing the sound from another L2. (Danish vs. Norwegian, er, jeg, hvis, osv.) (Unhelpfully, they also like to correct my pronunciation of non-Germanic words and names.)

Once I know my gaps and issues I can dive into understanding why. “Graag gedaan” was really hard for me because my mouth did not understand how to do the uvular trill and the uvular (French) r. To learn that, I did Googling about this specific aspect of Dutch Phonology.

Current me knows the answer, so it is easy for me to find all kinds of articles, videos and images explaining it with a little bit of Googling. The key is when I do not know the answer (and many times not even the question), I spend a lot of time trying to figure out how to explain to Google my problem.

Rarely am I the first person to have a question.