* First and foremost, Duolingo isn't a product, it's a series of products, which all have different characteristics. A person using Duolingo on Android will have an experience completely different from the experience of someone who uses iOS. And the difference is even more drastic when comparing mobile with the desktop version. The desktop version is sometimes three or four time more challenging.
* Even within one platform, the A/B testing has become so large-scale that even two Android users might see a completely different product.
* Courses for “popular” languages get way more attention than the “unpopular” ones. And the quality varies greatly.
* The ads they show are sometimes loud and obnoxious, NSFW, or straight up scams. That doesn't happen that often, thankfully, but I've had my portion of loud-as-hell game ads and borderline pornographic hentai game ads. And before you say what people always say when it comes to NSFW ads, Duolingo says that ads aren't personalised, so no, my search history has nothing to do with them.
* Finally, it is indeed a good way to learn some basics and acquire some basic vocabulary, but there is no way you'll get fluent with it. A friend of mine has summed Duolingo up very well: it gives you a feeling that you learn something, even if you don't progress at all.
That's just the stuff I could remember off the top of my head. It's still nice to have it, but we still should remember that the service has lots of issues, some of which could be eliminated, if the management wanted to do so.
Also a bitch to maintain with three experiments that might or might not be going on at the same time in one screen.
What I've realized is that experience teaches you to speak, read, and write a new language. You need a lot of experience to become good at it, because the language has to feel intuitive if you want to hold a conversation in it. There's the recommendation that if you want to learn a language you should immerse yourself in it (in real life). I think the reason why that works is because it forces you to figure out how to use the language and gives you an immense amount of experience in it.
I think that Duolingo is just another way to get slightly more experience in the language, but it's probably not going to be enough on its own. You're just not going to be spending hours every day on it to compete with language classes. From my experience, it was a nice way to quickly learn some of the basics of Japanese, but Duolingo really can't make you understand Kanji more easily than other methods.
I’ve learnt languages with immersion and Pimsleur. Sadly I haven’t found Duolingo useful at all. The effort to actual useful knowledge ratio is high.
Other helpful resources: Assimil, Anki, and Conversation exchange. Mundo Lingo is good for that if they have an event locally.
I also did it in parallel with the Language Transfer podcast, which I've mentioned on HN before but is an amazing resource for understanding how a new language works structurally and grammatically.
I did find Duolingo valuable, but purely for vocab.
https://static.duolingo.com/s3/DuolingoReport_Final.pdf
A few notes about it though:
1. It was funded by Duolingo but carried out by an outside team of academics. I can't judge how much this impacted their analysis.
2. No control group, which isn't very promising for the rest of the methodology.
3. More time spent using Duolingo did seem to result in higher improvement
4. Lot of dropouts, some participants excluded for taking up courses. They tried to control for 'outside resources' like watching movies in Spanish but it's unclear if that problem was really solved.
5. Novices of the language learned the most. Makes sense, as Duolingo focuses a lot of vocabulary and basic sentence construction
Personally I'm using Duoling to learn Italian, and I find that I am definitely learning, especially in terms of conversational phrases and vocabulary.
However, I've learned a few other languages by immersion, and this is going much, much slower. I'm 99% sure I would learn more by moving in with an Italian family or by consuming all my media in Italian, instead of Duolingo. For now, Duolingo is a nice compromise, and I accept that I'll learn a limited amount from it.
At least it saved me a hundred euros, but it's not a substitute for a real course and real interactions.
But later it talks helped in using more esoteric grammar that I never hear in small talk.
I'd suggest to use the website from time to time as it has some well done explanation to what is grammatically going on (the app is theory free)
Duolingo is a feel good tool - it makes you feel like your learning, perhaps quickly, without really learning anything.
For example, it uses multiple choice questions. Presented with a word (ananas), it then shows a few pictures (apples, bananas, pineapples and pears, for example). You incorrectly choose 'bananas', because really who wouldn't, and the pineapple lights up in green.
This method is really at odds with the 'best' ways to learn. If your recall is dependent on contextual cues, you will struggle away from that context.
Additionally, the volume really isn't that great. Like a lot of 'feel-good learning' platforms, duolingo encourages consistancy over volume. It's better to study for 10 minutes a day, than for 90 minutes every sunday. That's true, but if you want to learn a language (no small feat), then even 90 minutes a week is woefully inadequate. Being conversant in multiple languages is something to be proud of, and things to be proud of tend to take substantial effort.
Vocabulary acquisition is almost the perfect problem for spaced repetition, so use anki. There's a few strategies here: english word -> foreign word, picture -> foreign word, foreign word -> foreign synonym, foreign word -> english options, english sentence -> foreign sentence.
Pronunciation is difficult at first but gets easier. Accents are tricky - they're verbal gymnastics. Its easiest if you can find a native speaker of your target language, and you should just converse with them. Don't focus on minor details, because you usually pick these up naturally over time.
Listening is easy - listen to their radio, watch their TV. Similar with reading.
But there's one thing I haven't mentioned. Grammar. You do not need to learn foreign language grammar. You'll pick it up. How often have you studied English grammar? Unless you went out of your way, probably never. You learned some basic rules in your first few years of school, but everything else you picked up.
I had no Spanish background before Duolingo, or any spoken language other than English. I took Latin in HS, which helps somewhat with vocab and reading but not at all with speaking and listening. With solely Duolingo plus the additional practices I mention, I've gotten to the level where I can survive in a Spanish speaking country. I can't speak or listen well, my grammar isn't good, but I know enough to ask for what I want and understand basic information being told to me. I can also read enough to understand at least the key information from most written sources. I think I'll definitely need a more dedicated class and/or a longer full immersion experience to get towards my goal of being conversationally fluent, but Duolingo has been a great starting point.
It's pretty good for the first steps, and to get some vocabulary. I can't follow a non-basic conversation, but I can pick up words from native speakers.
I'm now at the point where I should work on my own with more seriousness.
Overall, it's good to start learning a language, and to keep a regular practice, but it's not enough on its own to actually speak the language.
I find that Duolingo is the least useful of these apps. Its main advantage is that it's easy and a small commitment if you decide to do one or two lessons every day. For the rest I don't understand why it's so popular. Duolingo's main strength is that it makes you actually construct the language instead of just memorizing individual words or expressions but I'm not super impressed with that either, mainly because in my experience it's very, very common to have a sentence rejected because it didn't match the internal "regexp" used to validate it even though it's perfectly correct (and sometimes these mistakes linger for literally years despite being reported). I actually got a very mediocre result in the placement test for the "French for English speaker" tree despite being a native French speaker who's reasonably fluent in English, mainly because some perfectly correct answers were rejected by the system. That really destroys your confidence when that happens (both in yourself and in the app).
You have a forum to discuss these issue which contains some very valuable information but it's the worst forum software I've ever used bar none, it feels like a teenager's first PHP project in the early 2000's.
Oh and if you use the mobile app it'll have you build the language by selecting one full word at a time, which means that you usually won't have to thing about the conjugations/declensions and just vaguely remember what word means what. It's fine for English, not so much for Russian and its complex declensions and aspect system.
For vocabulary Anki and Memrise easily win because they only do that and they do it well. You can find podcasts aimed at learners of many languages to improve oral comprehension and LingQ is great to improve reading comprehension, although it's expensive for what it is and I feel like you could make a better clone of it in one weekend.
In general I haven't been really impressed by any of these language learning apps, it seems that they really lack the resources to do anything but the bare minimum. It's probably too niche to generate some real R&D.
So overall if you enjoy Duolingo then stick with it, but keep in mind that you probably won't get anywhere just using this app. If I had to recommend only one language learning app it would be Memrise because it's got some decent decks for many languages (including user-contributed ones) although of course you can't learn a language solely by memorizing the dictionary.