I have been going through a textbook, Japanese for Everyone, and have been adding to two anki decks - a sentence deck with most of the sentences from each chapter, and a vocabulary deck with new words.

Currently I'm only doing recognition (in both decks), with very minimal kanji in the sentences and vocab at the moment as kanji are introduced slowly in this textbook (I have already finished RTK though so I don't mind).

But I feel as though there are sentences which I can understand when reading, but I would have more difficulty creating them myself.

So I am wondering, will doing only recognition cards disadvantage me in the long run? Should I make the switch to recognition and production now, so that I get used to it? But then there's the issue of multiple translations of the same english sentence. But, what do you think?

Recognition and production are very different skill sets, and training one will not really train the other. Some people prefer to focus on one skill at a time, but if you want well-rounded knowledge of the language it's advisable to learn to write and speak simultaneously with your study of reading and listening.

I would strongly recommend against recognition-only. I tried it with Spanish (a much easier language than Japanese!) and I didn't get nearly as much mileage out of it as I get using production. I would, in fact, recommend production-only over recognition-only, but having both is often best.

In fact, for Japanese I use a system with three cards per fact: English->kana, kana->kanji, and kanji->English. When doing kanji->English I also require myself to give the readings of the kanji. Cumbersome? Yes, it is. But I don't think it would be more effective any other way.

Aijin Wrote:Recognition and production are very different skill sets, and training one will not really train the other.

On the contrary, I've found that production does significantly aid recognition. I don't do this with Japanese too much, though, because with that language I usually need to do both. With, say, Spanish, though, using production-only often works. (I do tend to use recognition anyway, though, except when it's a sentence whose meaning is obvious through context or cognates.)

furrykef Wrote:I would strongly recommend against recognition-only. I tried it with Spanish (a much easier language than Japanese!) and I didn't get nearly as much mileage out of it as I get using production. I would, in fact, recommend production-only over recognition-only, but having both is often best.

In fact, for Japanese I use a system with three cards per fact: English->kana, kana->kanji, and kanji->English. When doing kanji->English I also require myself to give the readings of the kanji. Cumbersome? Yes, it is. But I don't think it would be more effective any other way.

Aijin Wrote:Recognition and production are very different skill sets, and training one will not really train the other.

On the contrary, I've found that production does significantly aid recognition. I don't do this with Japanese too much, though, because with that language I usually need to do both. With, say, Spanish, though, using production-only often works. (I do tend to use recognition anyway, though, except when it's a sentence whose meaning is obvious through context or cognates.)

I agree with this, for me my reading/listening skills have become high but my speaking/writing skills are so behind. I did a lot of recognition cards, sure I can write kanji but from memory? not all of them. One skill does help the other but eventually, just working on production helps production, reading=reading,speaking=speaking and so forth.

vinniram Wrote:I have been going through a textbook, Japanese for Everyone, and have been adding to two anki decks - a sentence deck with most of the sentences from each chapter, and a vocabulary deck with new words.

Currently I'm only doing recognition (in both decks), with very minimal kanji in the sentences and vocab at the moment as kanji are introduced slowly in this textbook (I have already finished RTK though so I don't mind).

But I feel as though there are sentences which I can understand when reading, but I would have more difficulty creating them myself.

So I am wondering, will doing only recognition cards disadvantage me in the long run? Should I make the switch to recognition and production now, so that I get used to it? But then there's the issue of multiple translations of the same english sentence. But, what do you think?

Personally from my experience it does help other skills out but not a huge difference. Knowing to read/listen/understand will reduce the time you need to speak/write correctly but you have to get your writing/speaking up to speed. For me i need to work on speaking/writing up to speed and all.But lucky i still got a lot of time, If it really takes me 3-5 years to get to fluency i wouldn't mind at all.
(It's almost been 1 year and so far my input skills are awesome so far but output not even close lol. So i'd estimate another year to get it up to speed and another to just use the language in general) 3 years sounds right to me

I find that while flashcards are great for recognition, they're not good for production outside of vocab. Practicing creating sentences is way better with actually speaking to people, writing a diary (lang-8) or doing textbook style drills / activities.

Saying this, if you are on a basic textbook then production flashcards are okay, it's just when you know various ways to say the same thing it starts being a pain.

furrykef Wrote:In fact, for Japanese I use a system with three cards per fact: English->kana, kana->kanji, and kanji->English. When doing kanji->English I also require myself to give the readings of the kanji. Cumbersome? Yes, it is. But I don't think it would be more effective any other way.

Sounds like a bad idea. Not because there's a lot of production, but because there's a lot of english. For example, how can you possibly go english->kana? There's usually at least 3 japanese words for an english word.

The trap when using a system like this is that you become a translation robot. When you want to say something in Japanese, you think of it in English, then use the english->kana knowledge to say it in Japanese, which is a bad way of outputting Japanese since it leads to unnatural sentences.

caivano Wrote:I find that while flashcards are great for recognition, they're not good for production outside of vocab. Practicing creating sentences is way better with actually speaking to people, writing a diary (lang-8) or doing textbook style drills / activities.

Saying this, if you are on a basic textbook then production flashcards are okay, it's just when you know various ways to say the same thing it starts being a pain.

While I agree with you I do actually think doing production sentences helps. Particularly if you do a lot of them focusing on a particular grammar point, it really sinks in as to how its used. Either that or I think a good idea is to take a basic sentence and have all the different ways it can possible be said to get you very familiar in the way the language works.

It is inadvisable to do recognition-only for all of your studies. But personally I find SRS production cards to be too exhausting, so now I do only recognition cards in Anki, and practice production outside of a flashcard-review system.

Can someone give examples of production SRS?

jpkuelho Wrote:Can someone give examples of production SRS?

There are various ideas how to go about doing it, depending on what you want to learn to produce. The few production cards I have are aimed at producing the kanji for a word. What I put on the Q side is a translation (usually several words) and the kana, then I simply produce the kanji for it. I do this rarely because it's much slower and annoying than recognition, but it definitely helps with handwriting.

Icecream talked about an idea she had (i think) where you produce dialogue. On the Q side, you have a part of a dialogue with audio, then you're supposed to produce the correct answer. Personally, I don't like this way of doing it much since it seems like what you're really doing is memorizing a certain answer from a certain specific context, you're not actually producing anything.

Obviously recognition + production is better than recognition only.
But time is limited, and we all have to choose.
And production can take as much as 3 times recognition.
My choice? Master recognition first. To master Japanese recognition (reading + listening) will take many years. Obviously I will try some output every now and then. But in general, first input and then output.