I think Duolingo is great for forming habits. Ultimately, if you want to learn something new you need to work on that skill over a decent period of time, and this becomes hard to do without a regular habit.

Duolingo is also great at exposing yourself to the language gradually, but unless your goal is to simply be able to hold a conversation or say a few phrases, I don't think it will get you anywhere in the long run.

As I'm getting more into language learning (specifically Japanese), Krashen's input hypothesis seems to be an important part of the process. https://www.youtube.com/watch?v=NiTsduRreug Essentially he makes the distinction between "learning" that word A in language X has some meaning Y, and acquiring an unconscious understanding that word A means Y.

His main point being that this unconscious acquisition (that is, being able to hear a words/phrase and not having to put conscious effort into to translating it in your head) only comes from comprehensible input, that is, listening to the language and understanding messages (I'm not so convinced that it has to be entirely 100% comprehensible, rather from experience think language can be acquired even from something like 20% comprehensible material).

Communities have been rising up around Krashen's theory, one being AJATT (All japanese all the time). alljapaneseallthetime.com

The creator of the site advocated listening to native Japanese material 24/7 (or at least as much as you possibly can during the day) as well as making i+1 sentence (linking back to Krashen) flash cards using Anki (or equivalent SRS system).

More recent innovations on AJATT have been arising thanks to MattVsJapan (https://www.youtube.com/MATTvsJapan). He is currently developing a method he's calling MIA (Mass Immersion Approach) that refines AJATT into a more general approach that is applicable to any language (for now it's more focused on Japanese since that's what Matt is most experience in).

One common criticism of these methods is that they advocate NOT speaking until you are have achieved some form of basic fluency. The reasoning being that training your unconscious model on listening will prevent you from making mistakes early on, until you are ready to train your model from your own speech.

Honestly I'm not doing these methods enough justice, there is a lot of background and theory behind them, and MIA itself is still evolving (Matt is actively trying to change the way people learn languages, or if not that, at least the way people think about how language learning works). If anyone is serious about learning a language to a highly proficient level, ditch duo lingo, textbooks and tutors and read up on these methods instead, you won't regret it.

user:	risos
created:	December 25, 2015
karma:	35