Hacker News new | past | comments | ask | show | jobs | submit login

This comment would be 60 times more helpful if in addition to your strong opinion on the failures of learning with Duolingo it’d supply some of the good alternatives.



As someone learning Japanese I'm really appreciating tools built for JP specifically: Renshuu and Wanikani. Both use SRS (same as duolingo) but spend a considerable amount of time actually teaching the grammar and nuances, they both avoid starting from everyday phrases like "I would like sushi" to instead build a foundation first, and many other little things that make it a much nicer experience than Duolingo who's trying to use a very generic approach that maximises small term satisfaction in exchange for painful long term learning.


I was under the (possibly incorrect) impression that Renshuu was very beginner unfriendly and WaniKani skips the most basic stuff (hiragana et al) and is “just” to learn kanji which ofc is important. Was I wrong?


On WaniKani: that’s correct. In their FAQ (I think?) they link out to an article on Tofugu (aiui run by the same people) which gives you a couple good anki decks to learn hiragana and katakana. I started wanikani without knowing either, and found it manageable at the start by referring back to a hiragana chart. At some point I went through the decks, and after about two weeks I could read hiragana well enough to leave them behind.

Certainly not a complete resource for learning the language, but very effective for learning (to read) the kanji.


I’ve always wanted to learn Japanese, thanks for the tips!


I have started learning japanese roughly 6 months ago, and I luckily stumbled upon this fantastic Anki deck: https://ankiweb.net/shared/info/911122782

It focuses on teaching grammar and vocabulary through listening comprehension. The creator has put an immense amount of effort into it, to a point where I cannot believe its free. I highly recommend it.


If you are interested I created an open source web app for learning kanji with SRS using open dictionary data.

https://shodoku.app/

I am a year into learning Japanese my self, and kind of wanted to learn vocab and kanji at the same time (and also see example sentences for the vocab which I can put into my anki deck). My method is when I start a new kanji I pick a few words that contain that kanji, bookmark them (and maybe add to my anki deck), and then when it is time for a reading review if I can remember how to pronounce those word I rate it as good.


Thanks for sharing, I'll check it out. I am currently using Wanikani +Tsurukame to learn Kanji, from your description your approach sounds similar with more customization?

I just had a look at it, love that it also teaches the stroke order, this is something I have no tool for at the moment.


If you are happy with WaniKani, you should probably just keep using WaniKani. It is a fine app (though a bit pricey). But I talked about the difference a bit in another thread https://news.ycombinator.com/item?id=43839879

The gist of it is that I like studying vocab and components (radicals) at the same time as the kanji. I kind of swap out the on/kun-yomi reading practice with bookmarked vocab, if I can remember the pronunciation of a couple of words with the kanji, I mark it as good. I also think writing the kanji helps remembering it (although I‘m not strict about it; personally if I screw up the stroke order, or add an extra tail, etc. I still rate it as good). I am also a fan of self rating rather than input evaluation.


Did you create an account just to shill this?


My learning finally picked up speed again when I started using CCI (Compelling Comprehensive Input). How easy it is to find material differs a lot between languages. Way way back in time I learned English that way, though I didn't think of it as "learning" back then - I was so focused on what is now called "compelling input".

However, you'll need some kind of foundation, otherwise it'll be hard to find anything to start with. Though at the language school my wife attended the teachers had methods for that too, when there weren't any common language to "teach" in. Show and tell, basically. Point down and say "This is a table". Point away and say "That is a window". And so on. The Krashen initial method basically, though the one teacher I talked to had never heard about the guy.

When I started Japanese I didn't use textbooks or classes, I used an app called "Human Japanese", which teaches structure and a little grammar, but mostly through show and tell. No conjugation tables or other boring stuff. It quickly gives you enough to start acquiring other material. My own huge mistake was to switch to Duolingo.


I'm trying to look up what this CCI thing is, but I don't seem to get further than simply "use the language". Do you have a good resource that explains how to apply the method or, if applicable, an example of a CCI course?


That's basically all it is. Find compelling content that is basically comprehensible at your level and consume it. You'll need to translate it, but you don't need that to be rigorous or to perfectly understand it. As long as you comprehend it in the end, you'll start to learn the patterns of the language subconsciously.

It's essentially a move from memorization and learning language as if its math to how children naturally learn language through exposure.


Yeah, I really don’t get all the hate towards DuoLingo on this site. Granted, it isn’t going to make you fluent alone but it is very good at keeping you sharp and getting your feet wet.

Name one sole app/course which will teach you absolutely everything there is to know about a given subject. There are none. All learning needs multiple avenues in order to be effective.

Even if you take part in a course with tutors they will you to practice out of the course and in your own time. Personally I found DuoLingo to be extremely helpful in getting the basics of Hindi down.


Because duolingo is designed for addiction (that's how they make money), not actual learning (learning would mean you'd stop using the thing, no good for stakeholders).

There is no sole app that makes you go from 0 to C2, but there are infinitely superior tools that actually make you learn, and not the self-complacent pretend-like-learning pastime that duo is.

For a start, almost every other app succeeds at not treating you like a toddler and not resorting to emotional manipulation.


I have to disagree in that you would stop using the app if you learn a language. Learning is a lifelong task and becoming proficient in a language does not mean you will stay proficient in a language. It takes constant refreshing in order to keep sharp.

Is Duo the best thing on the planet? No, does it serve a purpose? Yes. The reality is that, if people see their skills improving as a result of using the app (gamification etc included) then it doing its job.

> There is no sole app that makes you go from 0 to C2, but there are infinitely superior tools that actually make you learn, and not the self-complacent pretend-like-learning pastime that duo is.

This I strongly disagree with. Nothing can _make_ you learn other than your own willingness to do so. If you have the desire to learn, you will. If you do not, you won't. It is that simple and that is applicable to any subject.


> Learning is a lifelong task and becoming proficient in a language does not mean you will stay proficient in a language

Agreed, but most people see it anyways as a journey from point A to point B, and then it's done. Also, most people just settles for good enough, not continuously improving.

> if people see their skills improving as a result of using the app then it doing its job.

Problem being that duo tricks you into believing you are learning when you indeed are not. I feel encouraged when I understand something for first time, not when the godam owl gives me a high five because I matched a word with a picture.

> Nothing can _make_ you learn other than your own willingness to do so

Well, I am really willing to be a world class piano concertist and astronaut. Doesn't mean I'll become one. Motivation + habits set the baseline, the mimimum needed, but they are not remotely enough. Success would be pretty darn easy then.


> Because duolingo is designed for addiction

For people who have trouble keeping up hobbies, that's a feature. Even if duolingo isn't the ideal way to learn, it's a lot better than something I give up on or forget about after a week.


Looking to apps to learn language outside of spaced repetition and talking to someone over video seems pretty naïve to begin with.


There are a lot of things an app can do for you. Spaced repetition is the easiest one. However there are a lot of other options if you get creative. Most of them are a lot more work though. (though chatbots should now be easy as well to implement)


I know any of us can demo a feature, but how specifically is an LLM gonna correct your tone when it can't understand tone itself?


I agree i can speak passable spanish with my wife's family. i learnt exclusively on duolingo.

I don't know if its the best way but it kept me motivated to come back and put in some work in a fun environment. which i belive is the biggest problem to solve for any sort of learning.


The suggestion is that it's likely that you did much of your learning from speaking to your wife's family, with duolingo giving you a kickstart and the confidence to do so.

Having conversation partner(s) to practice with generally trumps any other learning method for languages.


I agree, for me Duolingo was great to learn the basics of Spanish, enough so that I could move on and practice in real life.


I think the pre-internet ways are just fine - textbooks, phrasebooks, other kinds of books geared towards self-learners.

With them, one must be just a little bit more proactive, though.

You can also sign up to in-person classes.


Pre-internet ways failed to teach language super often. Very frequent issue when learning from book was that you could not not understand anything people say, because you imagined the language to sound much differently then it does for months and months while learning. That was the most common result of language learning attempts - not much.

Language learning is one of the things that were genuinely made much more effective by the internet and streaming services. The input based learning methods were basically impossible pre-internet for most people. And these are very effective.


Many language learning books used to come with audio media. I'm old enough to own a few that came with cassette tapes.

Books are still worthwhile IMO, if only because they provide a bit of structure to one's learning. With free resources it's way too easy to become paralyzed by choice.


I am old enough to remember them. Comparably, you got maybe 4 hours of media - meaning sentences from the book being read and short boring dialogs. You cant compare it to what is currently available. It is like comparing a puddle of mud to Atlantic Ocean. And I mean it in a positive way - those audio tapes were almost nothing comparably.

Beyond projects like Dreaming Spanish, you have around infinite amount of French, Italian, Spanish or German Youtube about whatever topic you want to. There are even dedicated playlists for total beginners you can start to consume with zero knowledge. You have thousands of shows on Netflix in foreign language with various difficulty - some actually suitable for beginners. Some you have already seen in own language, so you can understand them more easily.

For major languages, there are dozens if not hundreds of podcasts with simplified news, "for beginner" discussions. Some of them are useable with literally miniscule amount of knowledge.


I think books are probably the worst way to learn a language

I learned French and my experience from best to worst ways to learn were:

1. 1-1 lessons with language teacher (by far the most effective way to learn)

2. audio lessons (Michel Thomas Method)

3. Visiting France a lot, interacting with French people (my wife is french) (and yes, for me this was less impactful than listening to audio lessons)

4. Duolingo (did a year of doing it daily, did almost nothing for me except a bit of vocab)

5. School (3 years of French in school was about equivalent to listening to 5 hours of Michel Thomas audio lessons)


Schools vary a lot. Some schools are really good, but a lot of them are bad. Schools are typically held back by those who don't care and so disrupt the class. (this isn't always bad, for kids learning how to deal with other kids is itself an important lesson - home school kids tend to do well on tests with much less time spent in study, but they always show a lack of meeting diverse people in my experience)


Italki, LingQ, Languagetransfer, StoryLearning...




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: