I consider myself pretty fluent. (Passed JLPT 1kyuu, can read novels with a high degree of understanding but not that quickly -- am still missing a lot of scholarly or subject-specific vocabulary.) I go through periods of doing SRS and not doing SRS, but either way the bulk of my Japanese-studying time is spent in extensive reading/listening -- more as entertainment than as studying, basically.

This is the way that I'm coming to think about SRS:
There are some words that are so common that you're going to learn them solidly if you get a good amount of exposure (though this is highly dependent on what you get exposure TO.)

There are some words that are rare enough that you might read ten novels and only see them once, or twice. When I SRS these words -- these words I've seen only a tiny handful of times in my life -- I don't really feel that I'm learning anything. No matter how many repetitions I do, it's just memorization, it's not language learning. I don't know if that makes sense. But the word doesn't seem like it's mine until I've seen it more times in natural contexts. I once though that SRS could act as a kind of stopgap until that happened, but it doesn't really work for me that way -- not if the gap might be months or years.

Then there are the words in the middle. Words that you've seen a handful of times, and you sort of kind of know them but not really. For THOSE, I think that SRS is quite useful, just to get it over the threshold of "I've seen this enough so that I know it." But I think almost any kind of conscious engagement with the word would be useful, even just writing it down along with the definition. 

Once you're at an advanced level, by the way, I think the words that you need get a lot more dependent on genre and content, so the more advanced you get the more sense it makes to do sentence-mining or take vocabulary words from authentic material, rather than just studying a premade deck.

Taishi wrote:

…but I usually make the comparison that 10 minutes of Anki in one day is clearly better than 10 minutes of immersion, but I can't really say 10 hours of Anki in a day would be better than 10 hours of immersion. At least not to the same degree, and it would be horribly boring. So why not take 10 minutes of Anki and 9 hours and 50 minutes of immersion. That way you get the small boost from Anki while not really taking up much time and effort.

Ah~, I see more clearly now. Thank you both!

Last edited by Marumaru (2013 January 14, 7:52 pm)

In all fairness to Fillanzea she has spent ages in Japan and many people on the forum don't have a chance to go

I don't know how you define "ages" but I've spent less than a year in Japan. It's really easy to download TV and movies these days; it's really easy to mail-order books; the human interaction part is a lot tougher if you're not living in Japan, but heck, I was a little bit of a hikikomori back then...

HonyakuJoshua wrote:

I think you get more "fluency" from oral aural input and practice than SRS, but still think srs is important.

I agree with this, SRS is good for stuff that needs memorizing, but fluency comes from getting the "feel" of the language. An 8 year old boy is still fluent in his own language even if his vocab isn't all that large.

That said, as an adult, you still need that vocab as well.

Last edited by Taishi (2013 January 14, 8:43 pm)

I haven't studied Japanese actively for about a year now though I still use it in my everyday life. I used to SRS every day and built a massive vocab deck of 18,000 words.

I definitely notice that I can't read aswell as I used to but that's because I don't have too much time to read Japanese like I used to. I think speaking is still just as it's always been, same with listening. So... I guess if I had of kept up decent amounts of reading that'd still be where it was.

Once you get really fluent, so long as you actually use the language all the time then you don't need to SRS I don't think. Though, I guess it still does help it becomes a question of effort to reward ratio and at the higher levels that reward starts to get rather small for such a big effort.