kfmfe04
Member
From: 台北
Registered: 2007-10-21
Posts: 482
Does anyone know the reasoning or the studies/evidence behind SRS Theory which says that having spacing between reviews is better than having many continuous reviews for the purposes of moving knowledge from ST memory to LT memory?
Is it just a matter of efficiency (ie either way would work)?
Or is it a matter of effectiveness (ie having spacing helps move knowledge from ST memory to LT memory better)?
Tobberoth
Member
From: Sweden
Registered: 2008-08-25
Posts: 3362
I can give you proof of effectiveness right here:
Make 5 flashcards with russian words. Russian word one side, English on other side. Review those flashcards 20 times a day, each, for 3 days. The end of that week, you will probably know those words really well. Don't touch those cards for 1 month and check again. Do you remember all of them? Doubtful.
Put 5 Russian->English facts into Anki. Just follow the standard pattern. Will you know all of them after 1 month? Yes.
Now let's see, how many times did you check every single card with the first method? 20 x 3 = 90 times per card.
As for anki? About 3 times, 4 maybe.
I mean, however the actual stickage to LT memory works in traditional flashcarding, SRS is obviously almost 20 times as effective.
Last edited by Tobberoth (2008 December 19, 3:12 am)
rich_f
Member
From: north carolina
Registered: 2007-07-12
Posts: 910
My opinion on "AJATT" is that when you build an immersion environment, you have more opportunities to encounter the material you're studying in a different context, and thereby reinforce your knowledge of it.
I've had this experience a number of times just from reading Japanese language stuff in general. I've even had it in my own Anki deck, when I'll add a second sentence, even sometimes unintentionally, that adds a second instance of a vocab word I may have been struggling with, but the second instance adds another point of context that makes the word suddenly "click" in my brain at a different level. I'm not really sure what's going on, all I know is that it works for me.
mattyjaddy
Member
From: Japan
Registered: 2008-11-15
Posts: 64
I think that SRS and language acquisition are separate beasts.
The SRS, as is mentioned in one of the wiki articles, is for lists of information. A language is not a list of information, so an SRS is not adequate for learning a language. But it's quite adequate for learning the kanji, which amounts to a list.
The reason SRS is good for sentences is because it allows you to encounter language that you know you'll likely understand a number of times. When you encounter language that you understand you are acquiring that language. Therefore, SRS'ing sentences you understand helps you learn a language. But as was mentioned, the most important part is creating environment where you will constantly be encountering the language. So, in my opinion, the use of SRS for sentences isn't really so much based on the spacing theory as it is upon the input hypothesis which says that input that you understand equals language acquisition.
mattyjaddy
Member
From: Japan
Registered: 2008-11-15
Posts: 64
Tobberoth, I don't hesitate to say that reading/hearing something that you understand 1000 times DOES make your Japanese better. Language is equal to knowledge/competence that is housed within your brain in the form of neural connections. Those connections are made stronger with every exposure. So maybe you "know" some part of the language, but hearing it again makes it stronger. Perhaps it makes it so that a certain grammatical aspect comes to your mouth .0001 seconds faster than before when you hear it that 1000th time than the 999th time. And there is absolutely NO doubt that there is a huge difference between hearing something you understand the first time and the 1000th time. But of course, the growth is limited to the grammar included in that one bit of language. As sutebun said, more and varied input is needed for more complete grammatical acquisition.
But then again, I wasn't really talking about reading the same sentence 1000 times. I was just talking about the fact that an SRS guarantees multiple exposures to language you are more likely to understand over a fairly short period of time.
Vosmiura, I understand what you are saying about the spacing, I think. But the spacing theory, I think, is perhaps most appropriate for conscious learning (how to write kanji, for instance). Acquisition is an unconscious process. Natural language acquisition doesn't rely on conscious learning. It relies on exposure to language you understand. So more exposure equals more acquisition, whether that exposure is spaced or not. But this is just a feeling, it could be an interesting thing to study to see if spacing has a significant effect on language acquisition. As I said before, the SRS works for sentences based on the simple fact that it guarantees repeated exposure to language you know you'll understand which is good for acquisition. It's especially good for Japanese since we can alter the input to be only in kana from which we must recall (write out) the kanji. This is important for making the writing of kanji automatic.
I guess we all agree about that, matty, but maybe focusing in something new might have a better effect. I'm very scared to stray away from SRS use, I see it as the great idea that it is, but even so it has its flaws.
In my view, you solve the problem your are pointing, by limiting your content to urgent material. Only put those things you really need to understand.
For us beginners it is easy, we need all the kanji for writing, most of the readings from all the joyo, a few more from outside the joyo (maybe a lot more), tons of vocabulary.
It is really hard to screw it up. The only thing you need to run from is from redundancy.
Last edited by mentat_kgs (2008 December 21, 11:31 am)
