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
21

“He is the perfect antidote to the purism you allude to which is unfortunately rife in online language learning circles.”
I’m not sure if “input purists” represent the majority of input-oriented learners.
Usually the question is simply: when do you start to speak (or write), i.e. right from the start / earlier or later?

Personally, I prefer speaking (and, if possible, writing) as early as possible because the level of depth of engagement with the L2 is higher compared to listening/reading alone or in combination.

Reg. Benny’s approach:
“He combined early output with phonetics training. […]
nailing pronunciation later which may work well for some languages but for mandarin tends to be a disaster.”
Definitely a good idea for tonal languages! However, at the beginning of my language learning journey in Japanese (and its pitch accent), I felt this was overkill. So it really depends on the phonetical distance and complexity of the L2 to be acquired.

“He noted down every correction and drilled it as an Anki flashcard to make sure he didn’t make the same mistake again.”
If I remember correctly, Benny himself is a huge fan of Anki. So an SRS is definitely something he uses in his language learning journeys!

If you want to dig deeper, here’s Benny’s current language hacking guide for Mandarine (published in 2022):
https://www.amazon.com/-/de/dp/1473674271/ref=sr_1_1?__mk_de_DE=ÅMÅŽÕÑ&keywords=Language+Hacking+Mandarin%3A&qid=1654682395&sr=8-1

By the way, I’m a bit puzzled (at least from a teaching perspective) by your conclusion:
“The exact reason why some outliers are able to acquire Chinese at lightning speed while most learners lumber along for years before achieving anything like fluency remains a mystery. The factors involved are varied and often too intangible to isolate with precision.”
As I wrote before, Will has found and combined established language learning techniques. There is nothing new or revolutionary here - and that applies to the methods individually as well as in their mix.

If learners then put in the hours of quality time (let’s say ca. 5 hours a day, ca. 6 days a week, ca. 1552 hours in a year), they will get results.
Some learners will progress a little bit faster (e.g. depending on their experience in SLA, their previous knowledge of related L2s, etc.), some will progress a little bit slower. However, there is nothing mysterious about making progress in processes of practical skills acquisition (the science of expertise/expert performance has made this point crystal clear, see, for example, “The Cambridge Handbook of Expertise and Expert Performance”).

In other words: Will has followed a deliberate practice-based feedback loop (input and output → corrections → sentence mining → SRSing → input and output, etc.) for probably 1500-2000 hours. And that’s the reason why he got good results.

In short:

  • Tell me which experiences in SLA you have,
  • tell me which SMART goals you have,
  • tell me what you want to do exactly,
  • and tell me how many hours of quality time you want to put into language learning every day / week
    then I’ll tell you what is possible in SLA - and what is not.

This process is only “mysterious” when

  • learners have no previous experience in independent SLA
  • have no SMART goals
  • have no established habits
  • have no real stats
  • and rely on feelings and beliefs.
    In sum:
    If Will has spent about 2000 hours acquiring Mandarin in 1.5 years (and has achieved a solid B2 level of speaking and listening), then he is more or less in the normal range. In other words, he is definitely not an outlier aka an SLA wizard (or something like that).

PS:
In this interview with Rita, Will said he studied Chinese 4 h a day and later switched to 2 h per day. He SELF STUDIED Native-Sounding Chinese in 1.5 Years?! How Did He Do It? - YouTube

So in 1.5 years he should be in the 1500-2000 hour range.

22

Reg. Benny’s approach:
“He combined early output with phonetics training. […]
nailing pronunciation later which may work well for some languages but for mandarin tends to be a disaster.”

  • I’m confused. That wasn’t Benny’s approach to Mandarin. Benny actually never achieved mastery of tones. He certainly didn’t take Will’s approach of nailing this from the start so I was citing this as as a significant difference between Will and Benny. Their approaches are only similar in the rudimentary sense that they both started speaking early. Other than that there are virtually no similarities.

By the way, I’m a bit puzzled (at least from a teaching perspective) by your conclusion:
“The exact reason why some outliers are able to acquire Chinese at lightning speed while most learners lumber along for years before achieving anything like fluency remains a mystery. The factors involved are varied and often too intangible to isolate with precision.”

  • That’s not my conclusion, it’s a banal truism. I mentioned as a reminder that - despite my breakdown - we will never know exactly why Will learned Mandarin faster than any other learner I’ve ever met or interviewed for my podcast.

In sum:
If Will has spent about 2000 hours acquiring Mandarin in 1.5 years (and has achieved a solid B2 level of speaking and listening), then he is more or less in the normal range. In other words, he is definitely not an outlier aka an SLA wizard (or something like that).

I don’t think his level according to the European Framework is relevant. The fact is regardless of how well he performed on an assessment his spoken covnersational Chinese is utterly phenomenal, qualitiatively in a different leangue to what most learners - including those who have been studying for decades - achieve, regardless of their performance on standardised tests. Of course this is something that can only be appreciated by Mandarin speakers.

I have sent his video to some of THE elite learners of east asian languages. Everyone I’ve interacted about this with reacted in astonishment.

23

Thanks for sharing, it’s always interesting to see the various approaches successful learners have taken. Especially if these are completely different from my own approach, for example I couldn’t bring myself to do SRS or study phonetics.
A couple of thoughts follow:

1
I have to say I’m a bit surprised by his use of the IPA, I always thought this was baloney :slight_smile: The same goes for Anki: but since it takes center stage in so many success stories - it must really work. So, it might be I who needs to rethink everything…

2
Will’s example is one among many that seem to disprove the Antimoon / AJATT maxim of not speaking early. This has always been counter-intuitive and I don’t see evidence that speaking early actually has long-term negative effects. On the contrary, it probably helps building a more meaningful relationship with the language by connecting with real people. Of course not everyone finds themselves in such propitious circumstances as Will did.

3
If you focus on the spoken language and keep the time you spend on the written language to a minimum, you will save an inordinate amount of time. I might have missed it but I assume Will did that? People like I, who want it all (at the same time), reading, listening, writing, speaking… are bound to be slowed down.

24

Thanks you for your response, you make some excellent points.

“If you focus on the spoken language and keep the time you spend on the written language to a minimum, you will save an inordinate amount of time. I might have missed it but I assume Will did that?”

Yes this was one of my main takeaways. The act of “spreading” - doing everything at the same time without prioratising a particular skill - may slow down progress a lot. He learned characters but mostly by reviewing flashcards on Anki so he could practice speaking and listening. He didn’t practice much reading. Everything was heavily skewed and focussed to spend as much time as possible with the oral language.

25

Definitely some impressive results, I understood nothing :). To be honest the method does not seem that novel, but another great example of the results people can have with the right blend of circumstances and application of a sound method. Some comments I have:

1
On IPA, I think this is a good way to help people when they can’t yet “hear” the differences between sounds. A way to get to that path. I think this is best applied to mined sentences, because words in isolation rarely sound the same in context. I believe Luca does something like this when beginning a new language.

We will never master anything without intention, and mastering phonetics is key to “sounding native”. If the goal is to understand and be understood, we can choose to not apply our focus here.

Just thinking about the people I interact with that are ESL there is something tangibly different about interacting with people that have “familiar” pronunciation relative to those that have a strong accent from their native language. It is hard for me to explain, but is very real nonetheless.

Last comment on this. Speaking in general is very weird. We are taking air from our lungs, and blowing it out holes in our faces while vibrating muscles and shaping our face holes (and muscles) in very specific ways. The IPA can help explain how we should be shaping those when it isn’t obvious, and allow us to train that muscle memory sooner.

I used IPA and images of mouth/tongue positions to learn a lot of sounds I had struggled to even understand how to make. A good example being the “French R”. The thing with muscle memory is it does not actually take that long to figure out, it just takes repetition over several days, and with speaking it pays dividends a hundredfold.

2
Input alone is not enough to learn to speak (well). A sound method is critical for learning, and learning is not a passive thing. It requires intention.

ANKI and sentence mining a carefully curated set of sentences are a great way to do this.

3
This is also hard to understate. I think it may be impossible to learn to speak a language to a high degree of fluency without regular spontaneous interactions in the “real world”. In general, I wonder what is the point if we won’t be building a social life to involve the language?

In this context I am only talking about speaking, I could imagine someone simply wants to learn to read for the sake of reading or engaging with the culture personally and that is totally valid to me.

But to learn to speak without anyone to communicate with, or no intention of connecting with others, seems both odd and impossible to me.

All of this is to say, the social aspect to me is key not only to facilitate learning to speak, but also part of the end result.

26

I’m not arguing that is is not useful for that purpose. It will definitely work. There are two use cases for sentences. One is as a model for grammar. Both glossika and sentence mining will do this. And the second use case is to memorize specific grammar/phrases/vocabulary for situations specific to you.

The second use case to repeat is definitely useful and yes putting that into anki will definitely work for some people.

For me, though, if I run into a piece of vocabulary or phrase in a conversation with somebody and it’s brand new I remember it based on the conversation and I don’t think I personally would need sentences.

For example: I did not know that the two Russian words for “now” had a subtle difference between them. The first is “sitchas” and the second is “tepeer”.
I was using them interchangeably.
When I had a conversation with the one guy for kazakstan he said "I understand you but you should use “sitchas” for “now” and “tepeer” for “now [that something has happened previously]”.

Another example of that would be “se fueron” and “se fue” in Spanish.
I remember when I was learning conjugations. I was in Mexico and my girlfriend at the time had gone shopping with her friends. One of her other friends showed up at the hotel room and we had a short chat, part of which she asked me “Maria y los demas se fueron?” [Maria and the rest went somewhere?] I had no clue because at that time I could not produce past “simple” but only past perfect.
After about five minutes of trying to communicate in a different way I said “Maria y las otras chicas han ido al mall” [Maria and the rest have gone to the mall] and she said “aaah… no se dice asi, se dice Maria y los demas se fueron al mall?” [you don’t say it like that, you say ‘Maria and the rest went to the mall’]. CLICK! and that phrase is burned into my memory. I have tons of examples like that. I also have tons of examples from learning specific colloquial phrases from watching youtubers or watching Spanish telenovelas.

Getting back to the point. I wouldn’t put “se fueron” into anki because from then on I’d just know it because I learned it situationally.

That said, for me, I don’t do speaking from the start like dude [or benny lewis] is doing. I’m more like Steve Kaufman in that way where I wait till I’m ready. So by that time, I’ve already absorbed enough vocabulary that I don’t need this generally.

In the future though, maybe?
I can see the use of certain model phrases and I do in fact use model phrases to construct grammar models. But again for me I think I’ll get those from something like glossika. As far as specific not-grammar related phrases, I think I’ll just remember them. But to your point: could they be good value? Yes I think they could [for me in the case of grammar models I’m struggling with] and for others for their own specific situation.
Also, if you regularly have conversations with different people you’re definitely going to want to know your own specific phrases related to your own personal stories. I don’t have tons of conversations with Russian speakers so I don’t need it, but sure it would help if you did. For you in particular living in Germany I could see it being very useful.

Anyhow, different strokes for different folks.

Dude is not wrong.

27

@noxialisrex:
I’ve dipped my toes into IPA before but not really been able to tie it one to one to “how to make this specific sound” so I think I’ve not been looking at the right material. It has helped me validate my amateur understanding I noticed with Spanish (where people stuck their tongue out when saying ‘d’) that the tongue position is different, so I’m vaguely aware that e.g. a chinese ‘r’ is not our English ‘r’ and son on.

Anyhow, if you have used IPA successfully, can you recommend some resources or links please?

Also: yes, Luca Lampariello sounds epic in his foreign languages.

28
  1. IPA is legit but you don’t 100% need it if the language is close to your own. In the case of faraway languages, (my personal experience is Russian) I find I can’t even make out the individual sounds which makes it difficult to remember them. Later, when I’ve had enough exposure I have the sounds burned in to my memory it becomes easier to remember the words. So I’m going to do IPA first (or something similar before I tackle any other distant languages in the future)
  2. Speaking early has pros and cons. If you’re crushed by embarrasment I wouldn’t try it. Being mocked by others (it happens) makes me give up. That said, if you have a friendly [and patient] conversation partner like a girlfriend/boyfriend or classmates it will help massively. There is also the risk that if you don’t do something like IPA you could burn in horrible pronunciation. So I think Will’s lucky choice of IPA was crucial for him.
  3. This is a philosophical point I agree with 100%. My personal position is that babies can learn any language. A language is only “difficult” based on how far away it is from your native language or other languages you know. And there are different areas of difficulty. I believe that spoken language is 2x-3x easier to acquire than written language and that unless you are a very small part of the population you will struggle to acquire a language through reading without assistance. LingQ helps provide assistance with that but I have noticed there are a bunch of people on here who can understand their TL written language well but cannot understand the spoken language. I’m much less interested in the written language so I just drop it. I believe that removes a huge chunk of the difficulty in languages that are not just in a distant language family but also have a different writing system.
29

Agreed. Trying to do everything at once I believe splits the effort. And if “everything at once” is a bunch of difficult things at once it will make it that much harder.

In the case of writing systems: I bet that written vietnamese would be a bit easier to acquire than korean or japanese or mandarin or arabic just because it’s written in latin script and you’re already familiar with it.

Since I’m not particularly interested in reading languages I just drop it entirely, the same as the guy does and focus purely on spoken and video.

30

Right. That said I can hypothesis that without actually being engaged with native speakers it will still mirror what the FSI says: distant languages are doable but take more time.
In my own case (I don’t have any interaction with native speakers) I think I have made about 1/3 the progress I did with Russian as compared to French for the same amount of effort.

I think the key here is the guy is immersed and trying to actively use the language.

31

Yes. Hours is very important.

32

You’re right that his approach is novel. It’s still a combination of others.
Glossika alludes to IPA and so does the mimic method.

It does show as you say that IPA is super useful where the pronunciation of basic components is massively different from your native language.

And I agree with you there: Benny Lewis has made a huge mistake in thinking that just because you can learn a close language in six months that he doesn’t have to modify his approach to learn distant languages. The reason why he gets away with not doing something like IPA is that you can still be understood if you mispronounce some of the consosonants/vowels in German or Spanish. Not so in Chinese or other distant languages.

33

I do agree that IPA or something equivalent is likely to be a crucial difference when trying to master the speaking of a language distant from your own.

I also hypothesize that IPA in particular also helps with understanding what you’re hearing. I believe that if I had done IPA with Russian first then my acquisition of vocabulary would have sped up as compared with e.g. French.
The specific point I noticed is that I could learn 50-100 words of French per day comfortably whereas with Russian I got overwhelmed at 30 words per day.

Now, 11 months later the phonetic components are burned into my brain and I suspect I could learn 50 or more per day.

I tried an experiment with Mandarin a few months back for about a month and I could only learn about 5 a day. My success rate for recall was as bad as 10%. I simply couldn’t tell what I was hearing for the majority of words so I had no frame of reference for recall. I suspect that muscle memory and actually saying the component sounds as per IPA will facilitate the ability to recall words.

EDIT: “we will never know exactly why Will learned Mandarin faster than any other learner I’ve ever met or interviewed for my podcast”

This piece I don’t agree with. We are discussing it here and some of us are attempting to draw generalizable conclusions by the use of reasoning.
You yourself have pointed out that the use of IPA is different. Correct. It is. A couple of others have used IPA and have epic pronunciation. Luca Lampariello and the mimic method guy are two examples. As opposed to “we will never know” it is perfectly reasonable to draw a hypothesis from reasonable analysis. You seem to get very attached to data points and so stuck in the weeds that you struggle to philosophize and come up with a generalizable theory from hypothesis. Peter is attached to his own theories, that is true, but he is also willing to explore in order to update his belief system.

I tend to agree with Peter that the time aspect is core. I also agree with your observation that IPA combined with early speaking helped will out.
I hypothesize that there is in fact five relevant points here: IPA, early speaking and IMMERSION combined with anki sentences and audio/video input.

For those of us who do not have the ability (or desire) to be immersed the legitimate question is this: Is IPA the most important or is it early speaking or is it a combination?

My own experience is relevant here. I did not do either early speaking nor did I do immersion nor did I do IPA (but I’m aware of it). I did do anki and audio/video input. I noticed that I had a hard time at the beginning with memorizing words. I suspect IPA would have helped (independently of speaking or not). My grammar also sucks probably because I only have passive grammar (from watching youtube) so maybe Will’s sentence method in anki works because he’s being forced to actively recall sentence patterns (even though Will himself thinks he ‘doesn’t do grammar’).

But Peter is alo correct. Time is relevant. Five minutes a day would not have worked. The guy was essentially in a Chinese speaking environment for a big chunk of his day PLUS his anki and his TV shows/youtube.

34

“Their approaches are only similar in the rudimentary sense that they both started speaking early. Other than that there are virtually no similarities.” (@Michilini)
Hmm, maybe I’m wrong here, but I’d say Will’s approach

  • is very similar to Benny’s “output early” method (i.e.: focus on content relevant to the conversation, focus on topics/sentences that are important and interesting to you, speak early, use an SRS, use various input sources such as TV shows, podcasts, etc. )
  • enhanced by an early focus on pronunciation à la Idahosa Ness (https://flow.mimicmethod.com/ - btw, his content is free now!).
    This isn’t meant to be disrespectful to Will, but both experienced language learners and teachers generally take the approach of combining different methods to language learning depending on (the distance of) the L2 and other constraints (goals, time, budget, etc.).

So, at least from my experience with language learning and teaching, the “novelty factor” of “90 per cent Benny + 10 per cent Idahosa” is rather zero :slight_smile:

35

Good point.
But Benny usually has only 90 days for achieving a B1 / B1-B2 level in an L2.

Anyway, I’m not sure if Benny 2.x still ignores pronunciation (see, for example: 🔴Live: PRONUNCIATION Language Learning Questions Answered today | Do you have any questions about pronunciation and foreign language accent reduction on your mind to ask Benny? If so, then come at midday EST / 9am PST /... | By Benny Lewis | Facebook).

36

There is also the fluent forever method. They also use IPA and something similar to anki.

37

I think Benny actually only achieves a B1. Conversational level where you understand each other’s life stories in a chat is much easier than being able to pick up some random material audio or written and understand it.

IMO Benny’s big mistake was to take his success and think that is all there is to it without researching what other polyglots do and modifying his method.

It’s also important to his marketing that it’s possible to do it in 3 months/6 months/whatever. If he tells folks that it’s really going to take a year and a half of intense effort, most folks will not even start.

38

“I mentioned as a reminder that - despite my breakdown - we will never know exactly why Will learned Mandarin faster than any other learner I’ve ever met or interviewed for my podcast.”
Since 01/01/2021 ,I have detailed stats [date, time, description of the activity, skill (speaking, listening, and so forth), hours / minutes per day / week / month / year, etc.) for all the languages I’m focusing on right now, esp. Portuguese, Japanese, and Spanish.

Without concrete stats from Will and other advanced Chinese learners, it’s impossible to say how fast he truly achieved a high level of conversational fluency compared to other learners.

However, as @bamboozled and you correctly pointed out, it definitely helped Will a lot to focus only on the conversational dimension while ignoring reading and writing activities.

This is definitely an important message for L2s with complex writing systems. And this was also one of the main insights from our lengthy comparisons between xxdb’s and the “ultrareading while listening” approach on LingQ a few weeks ago.

That being said, according to Will,

  • he studied 4 months intensely (ca. 4 hours a day) = ca. 480 hours and
  • then switched to ca. 2 hours a day (probably supplemented by watching TV, listening to podcasts, etc. in his down time) = ca. 850 hours.
    That’s still at least ca. 1330 hours focusing only on the conversational dimension. If we include his downtime, he’s probably in the range of 1500 - 2000 hours (nota bene: this has nothing to do with the CEFR. It’s simple math).

Therefore, Will may be great in everyday Chinese conversations (I can’t judge, but if advanced Mandarin learners say he is, I don’t doubt it), but that amount of time combined with such a narrow focus is neither a mystery that other learners/teachers can’t understand nor lightning fast!

In this context, the hours of quality time are crucial because “1.5 years” sounds fast, but “1500-2000 hours” is the range to be expected for a distant L2 in the oral dimension and a B2 / B2-C1 / C1 level.

In short, outliers would look different…

PS:
“including those who have been studying for decades”
“Decades and no concrete stats” are as bad as the combo
“years and no concrete stats” :slight_smile:

39

“I do agree that IPA or something equivalent is likely to be a crucial difference when trying to master the speaking of a language distant from your own.”
I agree. In retrospect, I should have done that especially in Japanese.
But I followed the common wisdom that says, “Don’t worry about pronunciation in Japanese, it’s soooooo easy for Indo-Europeans.”

Then I heard about pitch accent - and everything changed :slight_smile:

40

The definition of an outlier is: “a person or thing differing from all other members of a particular group or set.”

There’s no question Will is an outlier by the dictionary definition.

Even if you assume he spent 24 hours every day for 1.5 years to get to his level in the video his conversational skills are still qualitatively better than everyone I know who surpassed that number of hours studying Chinese.

I speak Mandarin and know a lot of Mandarin learners. His level of conversational fluency is really rare regardless of time spent on task.

The interesting question is why. I think it has a lot to do with the combination of efficient methods and techniques he used.