Hacker News Reader

Comment by sweetheart | Hacker News Reader

I run a language learning app (leerly.io) which focuses on teaching language through comprehensible input, which seems to be at least somewhat how LLwN is approaching the problem of language acquisition. For those interested in how to get the most out of language learning tools like LLwN, some tips which are backed up by the field of applied linguistics:

- Don't translate! If you do, do so very sparingly. It sounds counter-intuitive, but stopping to translate often will just slow you down. That's because...

- The most important thing is just experiencing the language. You need hundreds and hundreds of hours listening to the language to really start to acquire it. Comprehension is inevitable, just optimize for time spent listening/reading.

- Avoid speaking. This has shown to actually hinder the process of acquiring a language. Speaking is the natural result of having learned a language. You'll notice when you're ready to start speaking a little because you'll occasionally have thoughts in your target language. Until then, speaking practice is virtually useless.

Maybe these will help you, as they've definitely helped me learn Spanish. Buena suerte :)

Note: These tips are also only for people who want to learn a language to fluency. If you just want to learn enough to order at a restaurant, that's a different ballgame.

View on HN

"Avoid speaking" This is 100% a myth. The Defense Language Institute comprehensively disproved this in the mid-1970's.

I know this because it's discussed on the first day of classes at DLI as for why you need to learn to speak the language not just hear and read it. Because my job title was voice intercept operator.

Pre-1976 the Defense Language Institute did not score or rate students speaking ability because only reading and listening were considered mission critical skills.

Post-1976 it was considered a mission critical skill because it had such a dramatic effect on students final listening and reading abilities during the culminating Defense Language Proficiency Test (DLPT). Subsequently an Oral Proficiency Interview (OPI) was developed and is considered an integral part of the DLPT.

Note: This was all direct, primary research done at the Defense Language Institute with thousands of participants annually, so it was direct cause and effect experiment.

It's hard to find the specific research results summarizing this, but here's one example from that time period. https://psycnet.apa.org/record/1975-10458-001

Not according to the research I've done ¯\_(ツ)_/¯, or my own experiences.

EDIT: As I linked below, here is Krashen on the subject - http://www.sdkrashen.com/content/articles/down_with_forced_s...

Right but as has been pointed out elsewhere, the linked article does not say to avoid speaking. Instead it says to avoid a strong focus on forced speech in a classroom, which likely means the sorts of repetition exercises that many of us remember from adolescence. That is very different from not trying to use the language productively at all.

Right. Making students repeat and memorize sentences that are beyond their current ability is detrimental.

Think of it this way. If I gave you a paragraph in Turkish (or insert language X) to memorize and recite that's very different than giving you some greetings and responses to memorize and recite.

For the first there is no mental model for where the paragraph fits in or how it's constructed. In the second almost everyone has some experience learning greetings in a couple of languages just by exposure to pop culture, so even if you don't understand the specifics of how those phrases work you have some confidence in how to employ them and you'll generally feel somewhat confident in being able to break them down later if you need to because greetings done in short and simple language.

I was surprised to see that suggestion as well: every language course I've ever seen focuses on memorizing a (relatively) natural dialog and reciting it word for word from memory as its primary means of targeting fluency. I don't know about the research OP cites, but I've found that, for me, memorizing the dialogues as the courses suggest have been by far the most helpful way to advance in language acquisition. I will admit, though, that of the foreign languages I do speak, I speak them better than I understand them... so maybe there is something to this.

The DLI has a different set of motivations: bring as many analysts to a working level of proficiency within a fixed time period, and has a lot of other things that confound this analysis: it's totally ok for the DLI to wash out some percentage of trainees in each flight, they can pre-screen candidates into whatever language they think is a best fit depending on candidates latent aptitudes, etc.

There is a universe where both pieces of advice can be right; surely you can imagine "don't worry about speaking" advice might be reasonable for someone who is learning casually, has no hard time frame to learn, wants to learn in dribbles and maximally passively, for whom a partial fluency might be a reasonable endpoint, and the greatest risk is burnout from negative reinforcement of failure.

Those are all different claims than the original post was making though.

"Don't speak" is a myth and by not speaking you are hindering your own learning. Dr. Krashen's advice was taken out of context. It was not meant as general advice not to speak.

If you personally don't care and don't wish to speak that's fine, but don't try to promote that as advice on how to get the most out of studying a language.

There's a lot of wishful thinking around language learning. The idea that some part of it will "just happen" naturally is very appealing. The unfortunate reality is that you meed lots of practice at everything: speaking, listening, reading and writing. You won't improve significantly at any one of those skills without many hundreds of hours of practice.

Massive amounts of listening and reading is practice :)

Yes. It will make you better at listening and reading, but not at speaking.

Yes, like speaking will not make you better at comprehension, although comprehension is required to have any meaningful conversation. Speaking needs to be practiced, but input and understanding comes first.

I'm pretty sure the originally poster did not mean to never speak at all. But just not to worry about it for a while.

I don’t see any reason to delay practicing simple phrases like “Hello”, “How are you?”, etc., or practicing the pronunciation of the vocab that you’re learning. It takes years to be able to reliably comprehend naturalistic conversation. You can’t realistically delay speaking until that point.

You definitely can, but whether it's necessary to delay simple things like that, I'm not sure.

It's pretty simple to speak things that you've acquired (of course it'll be slow and awkward at first, but that's natural), I just don't think there's much point in trying to speak things you haven't already acquired.

I interpreted the GGP to mean something like: think about how you learn your first language -- you spend a ton of time listening before trying to speak. Don't focus on speaking so soon. I don't think that is so much "wishful thinking" as an interesting strategy to consider.

I'd do it for spanish, but I can't find star trek with (latam) spanish audio.

I don't think so. Babies are constantly making sounds and trying to speak even when they have a very poor grasp on the language.

I personally learn much more from fumbling around with a language, and letting fluent speakers correct me, than I do from listening to it - though both are helpful. I recently had the opportunity to watch a lot of Spanish TV with a couple native Spanish speakers and I learned a ton from it. Through the commercial breaks I'd be repeating phrases from the show and asking them about it - or telling them my theories about what was going on in the show to see if I was understanding it.

There is little in common between language acquisition in kids and adults. And I have been reading the literature, and you have one research group that says it is the same, another saying it is different. Find me one kid (maybe there is one, but you get the idea) who is growing up in a household in which the local language is spoken and does not sound native when they speak.

>There's a lot of wishful thinking around language learning. The idea that some part of it will "just happen" naturally is very appealing. The unfortunate reality is that you meed lots of practice at everything

Not really. In actual life tons of people have learned to speak a second language naturally without any deliberate practice, books, etc, just by immigrating to another place and picking it up making acquantainces, working (e.g. menial jobs), and so on.

> - Avoid speaking. This has shown to actually hinder the process of acquiring a language. Speaking is the natural result of having learned a language. You'll notice when you're ready to start speaking a little because you'll occasionally have thoughts in your target language. Until then, speaking practice is virtually useless.

This has to be the most counterproductive advice I've ever heard. You need to start speaking as soon as possible. There's no other trick to learning a language than forcing yourself to speak.

For some it's easier because they're less socially anxious. For others it will be more difficult. I was in the latter camp learning Danish. You have to make friends with your fear or you will forever be stuck in what many language learners refer to as a "quiet period". I was for a decade (!!!) If you don't start speaking you will forever have only an intellectual understanding of the language.

So speak. Please speak. Early and often. Babies sound things out early because they're trying to get a hold of it, the vocal contortions required.

Now a result I can believe in this vein would be: "that attempting to lean heavily on sentence construction and grammar exercises to induce speaking could have a negative effect on language acquisition" and that's what I believe Stephen Krashen is actually advocating against.

But that's a very specific claim as opposed to "avoid speaking".

Hey, if it works for you, that's great! It's not necessarily what the research I've done says, but I'm glad you found a strategy that works for you.

EDIT: You may appreciate this - http://www.sdkrashen.com/content/articles/down_with_forced_s...

The title even disagrees with your statement. And this research is entirely focused on student anxiety. "Forced output" is what the claim is against.

Even more specifically these studies are focused on beginning students (i.e. 1st term students), so it is even less applicable once you have some level of experience with the language.

There is even a section on "How to promote speaking fluency". He is 100% not advocating to avoid speaking, but it talking about how speaking is a lagging indicator of proficiency.

Speaking requires the integration of language skills, so it will always lag behind other language skills. That does not mean it should be avoided.

The Defense Language Institute (DLI) even formally recognizes in its graduation requirements. The standard is 2 Listening/2 Reading/1+ Speaking indicating that speaking skills will generally lag behind reading and listening skills by 1/2 a level.

You haven't really shown how speaking improves your understanding or comprehension of a language.

Of course you need to speak to get good at speaking, I don't think that's what the original poster meant. But rather that you should focus on comprehension initially,because there's not really much point to speaking to someone if you don't know what you want to say and cannot understand the response.

Language comprehension is a two way street. If someone thinks comprehension is a passive activity, then they're likely not fully comprehending. Comprehension requires active engagement.

When you're learning another language, there'll be countless times where you're like "ohhhh, I understand this." Then the moment you open your mouth, you'll get pretty directly told, "No, you completely misunderstood." You'll also encounter people who immediately correct your mistakes when speaking, whether they're grammatical or misusages of words.

As an example that people here might better grasp, learning a language without speaking is like learning a programming language without writing your own code. You might generally recognize the syntax and usage of various functions, but you won't truly grasp it and understand its nuances until you sit down, write, and accept that you'll make mistakes and identify them.

>Comprehension requires active engagement.

I agree, but active engagement does not require speaking or writing. Listening and reading are not necessarily passive. You don't need to know how to write fluently to be able to read fluently and comprehend something, and I don't need to say something to be able to understand it. Can mute people not understand spoken word as well as people who can speak?

>As an example that people here might better grasp, learning a language without speaking is like learning a programming language without writing your own code.

>You might generally recognize the syntax and usage of various functions, but you won't truly grasp it and understand its nuances until you sit down, write, and accept that you'll make mistakes and identify them.

You'd need someone to correct your mistakes for this to work. And that can help, sure, but you can definitely internalize the correct usage of phrases, etc, from reading and listening to it being done correctly as well. Doing something incorrectly (which you will most likely do without a highly developed model of how the language works in practice) will not help on it's own.

People who are mute still communicate. They're not merely listeners.

If you're not practicing output, you're not truly comprehending. Output, whether it be speech, writing, or signing, is how you get and demonstrate confirmation that you comprehend.

If you were speaking your native language to someone who understands it, but spoke back to you in their native language that you are learning, you can get the same confirmation. I've done this before and it can work quite well. So I'm not really convinced it's necessary for comprehension. It's necessary for being able to speak of course, I don't disagree with that.

This also implies you cannot understand a language without people who can speak it to you. You can learn to read books without reading them out to someone who tells you what they mean.

Honestly, this doesn't even really make sense considering my own experience of going from not comprehending stuff like shows and podcasts to being able to almost completely understand everything being said a lot of times, without any real practice speaking to people. Unless you want to claim that I'm not actually understanding anything at all :).

I'm curious, do you know of any research (or even just anecdotes) showing that speaking practice is bad? I've heard the claim before but never seen any evidence for it.

In my own experience, it feels like speaking helps improve my speaking a lot. Sure it doesn't help me pick up new phrases or grammar at all, but the first time I say something always comes out horrible, before getting progressively better as I say it more. And there's the skill of utilizing a limited vocabulary to communicate complex ideas. That's another thing that I feel has gotten better as I exercise it through speech.

I'll try to dig up the original study I found months ago, talking about correlation between speaking practice and language acquisition in Japanese students learning English. When controlling for method of study, the researchers noted that students who tended to speak more ended up doing _worse_ on final exams. This could potentially be explained by the tendency for early speakers to way over-focus on correctness, which slows the whole process.

In the meantime, this book by Krashen may be of interest! He touches on some of the same ideas, roughly. Notably, around page 100 or so, he starts to define what he would consider his ideal learning environment: http://www.sdkrashen.com/content/books/sl_acquisition_and_le...

> the researchers noted that students who tended to speak more ended up doing _worse_ on final exams

Surely those exams focus on literary analysis or advanced grammar structures.

Who is the better speaker of a language? The person who can carry simple (daily life topics) conversations fluently or the person who knows pedantic language constructs but can only utter them with significant delays and only speak in chunks of a few sentences?

Knowing a language and being able to speak it are different things. I agree you have to know it (understand grammar and vocabulary) before it makes sense to practice speaking it. Until you are forced to quickly think on your feet in another language your speaking ability won't progress much.

> students who tended to speak more ended up doing _worse_ on final exams.

It is possible that this simply tells us that the final exam was only or primarily a written one. I know it is not directly comparable but my children learnt Norwegian starting at age 3 1/2 (after they already spoke English) without any formal instruction at all, just by speaking with other children and the staff at the kindergarten.

Living in Japan, I can verify that there are two major types of English students here.

Group one learns to speak casually, and often.

Group two can be handed a Shakespeare text and answer any question about it.

Group one will have a completely natural and fluent conversation with any English speaker. They can write casually and be completely understood.

Group two will be able to identify grammatical mistakes on a test. Many are incapable of even rudimentary conversation. These people did well on college exams and are responsible for the horrible, worse-than-machine translations you see on signs and government reports in this country.

I'm also in Japan. While I don't have any formal English school experience, my experience with casually meeting people pretty closely matches this.

Though, I've actually met some group one folks who can't write very well, despite being extremely proficient in conversation!

I have a Japanese friend in a third group. He learned English to a fairly high level almost entirely by reading English on the internet and watching English shows/movies. He's very fluent at speaking as well, but according to his account, his speaking fluency took a lot of practice (obviously), but didn't come until after he could already almost fully understand and enjoy those things.

I disagree with nearly all of your advice and tips. And I'd be very surprised if they reflected actual scientific findings in linguistics.

The least objectionable is the "don't translate" tip but even there I think you've gotten it wrong - translating isn't the issue - avoiding the trap of getting bogged down in trying to translate _precisely_ or being a perfectionist. But you definitely have to maintain some understanding of the context in order to get any value from subsequent exposure.

I fully disagree that you can acquire language just by osmosis - that it's enough to just listen to a language for hundreds of hours. Do you really think that if I isolated myself with hundreds of hours of say Mandarin audio recordings, TV shows and newspapers and novels and spent 40 hours a week listening to the recordings or staring at written text, that in a few months I would acquire the language? I mean if this actually worked, learning second languages wouldn't even be seen as a challenge or task.

It takes thousands of hours of active "work" with a second language to acquire it (although the data is patchy), not hundreds of hours of passive "experiencing".

And avoiding speaking is terrible advice. It's been shown to provide a huge boost to language acquisition. I know people who can read and comprehend the spoken form of a second language at a reasonably high level (C2) but can barely speak it. Speaking will not just "happen" if you don't actively practice it.

If you honestly did manage to acquire fluent Spanish purely by just passively "experiencing" the language, then I'm guessing your first language is closely related one - one of the romance languages.

> Do you really think that if I isolated myself with hundreds of hours of say Mandarin audio recordings, TV shows and newspapers and novels and spent 40 hours a week listening to the recordings or staring at written text, that in a few months I would acquire the language?

Yes, it does work, according to various accounts I've seen, although 40 hours for a few months might not quite be enough (the usual figure is around 18 months of full-time immersion, and more is obviously better), and this is actually quite challenging to do for long periods of time (esp. without distracting yourself with your native language excessively) and that's why it can be a challenge.

>not hundreds of hours of passive "experiencing".

It's on the hour of thousands, not hundreds. There are people, especially in the refold community, who have reached fluency in languages very different from their native, with mostly input (active output is part of the equation, but it can be done later, after your comprehension is very high).

>Speaking will not just "happen" if you don't actively practice it.

Yes, you do need to practice speaking to get good at speaking. But from aforementioned accounts, this process can be quite fast after a high level of comprehension has been gained.

I can confirm that for myself, over the past half year or so, increasing the amount of input I get every day has greatly accelerated my japanese learning process, versus the previous year, where I got far fewer hours in (1, maybe 2 hours a day at most, versus my current 4+ hours every day), although I'm not fluent yet.

In just that time I went from barely being able to read, and only being able to understand simple beginner podcasts to being able to fully understand easier shows (stuff for adults is still mostly out of my reach), and understand most of all the podcasts I listen to, (as long as I'm actually focused on the audio, if I'm passively listening I don't understand as much). My reading hasn't improved as much, as I spend probably 80% of the time listening/watching shows versus reading, but I can read basic manga easily now, and I'm working on reading through some essays now, novels still seems massively difficult, but we'll see in another couple months.

I've found that many hours of audio input each day have greatly improved my ability to keep up with native-speed speaking, although for increasing my comprehension, intensive study (making sure I understand each line in a show, or careful reading) is more effective per unit of time. I have no doubt you could acquire the same results as intensive study with enough exposure to the language though, as I've seen it happen with myself before I tried adding intensive study, it would just take more time.

Most people wouldn't call 4 hours a day massive though, so I can probably still increase my rate of improvement if I find more time to use.

> Yes, it does work, according to various accounts I've seen, although 40 hours for a few months might not quite be enough (the usual figure is around 18 months of full-time immersion, and more is obviously better)

Could you provide a reference to these studies or accounts?

The original claim was a second language can be acquire by hundreds of hours of passive "experiencing". Your claim of 18 months full time is more like 6,000 hours if by full-time you mean all-day. This is more than an order of magnitude more than the original claim and 2 to 5 times more than then numbers I've seen in studies on how many hours it takes to achieve C1 level in a second language using regular teaching/learning techniques. There is significant variance depending on the the relationship between the learner's first language and the L2.

So even if it worked (and I'm dubious - learning anything requires some effort/active practice) spending 7,000 hours of passive "experiencing" would be a hugely inefficient way to learn a language.

Perhaps we are having some miscommunication here. I don't know if the original poster was making a precise statement when they said "hundreds and hundreds of hours" anyway.

>Could you provide a reference to these studies or accounts? https://www.youtube.com/watch?v=pxYZvMByRjU&list=PLT9cfjU1yk...

I haven't watched any of these videos in a while so I don't remember which ones are good, and it's a lot to go through, sorry about that. But there are some accounts here of people who have followed "mostly-input" style methods. I don't think any of them are 100% passive input though, and I'm not sure anybody argued that is the best way to do it.

I also seem to remember a study I found recently on whether uncomprehensible input works for language acquisition with very interesting results, but I cannot find it right now.

>spending 7,000 hours of passive "experiencing" would be a hugely inefficient way to learn a language

The efficiency really depends on if it's something you want to do or not, but I don't know if anybody is doing a full 12 hours a day (maybe some super hardcore people). I believe it's closer to 6, maybe 8 hours if you're very serious about it and have the time.

>learning anything requires some effort/active practice

Yes, you need to actually try to understand what you're listening to at least part of the time, maybe most of the time, which means you need to be actually interested in the content, otherwise you will get bored quickly. But passive listening while you are doing other things definitely helps in combination with active study.

I'm not sure if we're using "passive" in the same way though. Actively focusing on, and trying to understand what you are listening to isn't what I'd call passive.

Taking 4 hours out of my day is pretty massive. Between sleep, work, kid and basic household chores I wouldn't have 4 hours a day ever.

Yes, it also feels like a lot to me, but it has become much easier to do as my comprehension has gone up. But most of that is not actually taking time out of my day. I go on walks every day, listening to podcasts at the same time. While I do chores, browse the internet, or drive somewhere, I do the same thing. This usually adds up to at least 2 hours a day. On top of that I make sure to do some reading, then try to watch a show or two in japanese on top of that.

I suppose this last year has been quite good for immersion style learning, since there wasn't much else to do except stay inside.

Of course, those with less responsibilities can spend much more time doing this, and most of the success stories where people get truly excellent results in short amounts of time are from high school or college students. I don't mind if it takes a bit more time for me :).

Sure, don't translate, but the input also has to be comprehensible. So there is always a challenge of not seeing the translation but still understanding. If you never look something up, there are some words you will never know. There is a balance, and that balance is hard to find.

True, a lot of it depends on personal preference, too. I think the idea is to not worry about understanding 100%. Even just 70% could be fine! Or just the gist or emotion of the text could be enough.

This is wildly misleading advice. Using materials with comprehension below 80% is a massive waste of time.

Will you learn something? Yes probably, but will it be time and effort efficient? Not even slightly!

At 80-95% it can be productive if you're putting forth a lot of effort, but comprehensible input assumes comprehension above 95% for passive learning of materials.

To demonstrate what different (reading in this case) comprehension levels will get you take a look at this[1].

[1] https://www.sinosplice.com/life/archives/2016/08/25/what-80-...

Agreed, but this is only a problem is you are biting more than you can chew. You can either learn the unfamiliar words to increase the comprehension rate, or find a material that is more suitable to your comprehension level. Comprehensible input is for reinforcing and contextualizing the previously learned words.

A lot of people have already commented on the part of not speaking being virtually useless at the beginning. I tend to agree with the criticisms listed there, but I submit another reason: muscle memory.

Practicing making the sounds of the target language as soon as possible with the correction of a native speaker, in my opinion, is absolutely essential. For many languages, there are unique sounds that are just close enough to one’s primary language, that our muscle memory causes us to make the familiar, incorrect sound instead of the proper one. Because it’s more a matter of physical exercise alongside the rewiring of our brains to recognise said phonemes, the more time spent working on it, the better on the path to fluency.

Having said all that, everyone’s language acquisition ability and language goals will be different, so I can see where this piece of advice does not have universal application.

Thanks for the tips.

One thing I've noticed helps me learning language has been working hard to imagine the physical thing as opposed to the english equivalent word. That is imagining a shiny red apple when learning "manzana" instead of the word "apple". Always wondered if theres anything to back this sort of approach up or its just different learning styles.

Oh totally, I've wondered the exact same thing. I started messing around with that on leerly, and there are a couple articles with a feature called Storybook, where images appear as the user is speaking, to help build the association between the foreign word/sound with an image, rather than the associated word in your target language.

I have no idea if it's effective, I just think it'd be cool to see how it helps with learning grammar. I def want to experiment more with it, but it takes a fair amount of time to add images to all the articles.

Transover is a good tool for surgical translation:

- https://addons.mozilla.org/en-US/firefox/addon/transover-ff/

- https://chrome.google.com/webstore/detail/transover/aggiicla...

> The most important thing is just experiencing the language. You need hundreds and hundreds of hours listening to the language to really start to acquire it. Comprehension is inevitable, just optimize for time spent listening/reading

Listening with or without subtitles? I've been trying to figure out if watching content in another language with English subtitles actually helps me at all...

Nope, no subtitles in your mother language. Only your target language. It's okay if you don't understand a lot, what matters is exposing yourself to the sounds and sights of the language, and the general message of what is being conveyed ("I don't know exactly what they just said, but they seemed pissed and they said 'mierda' a lot, so maybe that means something bad").

Ideally, you'll want to focus on material where you can understand the majority of what is being said, through subtitles or just listening. If you find you can't do that, try watching something that would be easier to understand.

I couldn’t disagree more with “avoid speaking”. Being forced into conversations is my main method for learning another language, even with a small vocabulary it’s how I retain most of what I learn… by practicing it. I do agree with the “Don’t translate” though. That never works.

I started having thoughts in my target language after I started speaking. Until then it was just words on paper or on a screen.

wait until you dream...

Muchas gracias!

¡De nada, tío!

Reply on news.ycombinator.com