I've been reading a lot about learning recently, not in an ironic way but to clear my mind of some curiosity. One major bother for me has been understanding the effect of production vs. recognition and their role in language learning. Some material first...

http://www.visualthesaurus.com/cm/bookno...ocabulary/

Quote:Here are two nonsense words: balderate and mamiber. Let's imagine that all you know about these two words is that they have something to do with killing. If this is all you know about these words, you would not be able to use them properly in your own sentences. However, would you be able to understand sentences in which the words were used? See what you think:

a. Pierre Gemayel was balderated in 2006.
b. The soldiers mamibered hundreds of men, women, and children.

You have enough information to be in receptive control of these two "words," so you are able to understand the above sentences. If we give you the actual words (assassinate and massacre, respectively), you are able to tap into the fully specified information that you possess as a native speaker and use them correctly in sentences. In other words, you are now in productive control of these words.

As we stated earlier, acquiring productive control over words is normally a gradual process. Bits and pieces are added to your body of knowledge about a new word as you are exposed to it in a variety of contexts and situations.

On "holistic learning" (http://www.jwelford.demon.co.uk/brainwar...olist.html)

Quote:This powerful model presents learning as an interaction between four distinct modes of psychological being: feeling, imaginal, thinking and practical. These are normally represented in the form of a pyramid, as above, with feeling at the base and practical at the top. [...] The significance of this alternative orientation is that the crucial requirement for each learner is to establish a relationship with their total learning situation which is intimate, resonant and positive (i.e. in the feeling mode). Only when this is firmly in place is it considered that the learner will be free to tap fully into the other three modes of the learning model, viz. imaginal, thinking and practical.

Emphasis mine. It's obvious to most of us already that interest, self-esteem and motivation are fundamental to learning (the "feeling" base of the holistic model), but what I want to point out is that practical is placed at the very top.

These articles have me questioning the function of production cards as I go through Core (and grammar).

Say I'm an average 6 year old with "approximately 2,600 words of expressive vocabulary (words I can say) and 20,000–24,000 words of receptive vocabulary (words I understand)" (wikipedia source). While my school activities provide some production-based learning (quiz books, homework, etc.), the vast majority of my output is involuntarily acquired after I've gained sufficient receptive control of words to use them confidently. The vast majority of times I use a new word will probably be through my own volition to communicate. Pretty sure this is how I learned to use わたし[わ], それ/これ,　はい/いえ, おはよう[ございます], all the other words involved in 自己紹介, various curses/dirty words etc. etc. from the ~40 anime series I watched during the autumn/winter of 2012. Moreover, even after words become part of my expressive vocabulary, much of my vocabulary will still continue to develop more precision, tone, meaning and nuance as I read and listen to more and more material.

It's noted on wikipedia that "it is possible for the productive vocabulary to be larger than the receptive vocabulary, for example in a second-language learner who has learned words through study rather than exposure, and can produce them, but has difficulty recognizing them in conversation", which is fine, I suppose they have the same value. However, it leaves the learner in an awkward position when wanting to use productive vocabulary in contexts they have yet to see it used in, or even worse, associate incorrect contextual uses with a word* (not to mention lacking recognition). Is production vocabulary not better (if not naturally) acquired AFTER significant "receptive control"?

My point: I think that learning via production cards in the early stages (at the very least) of vocab SRSing has a negligible effect in practical terms, because you still have to wait for receptive vocabulary to catch up before having anything like enough confidence to use the word expressively. This leads me to believe that recognition cards offer much more bang for your buck, and it wouldn't be a bad thing to focus on them solely during core2k while listening-reading the sentences.

So when is a good time for production-based learning? I'm going to be hasty and say there is little-to-no place for vocabulary production cards in an SRS environment. That said, talking as much as one can in Japanese (to oneself or with someone else) and using Lang-8/blogs to practice writing are the best places to work on output, because that is what speaking and writing is for, where it happens, and the only time it has meaning (i.e. when trying to communicate). Reading and listening, however, have no practical limitations.

This is from another thread, which I can relate to as a mental barrier preventing me from trusting recognition cards by themselves, without the reassuring expectation of production ability (emphasis mine):

devilsbabe Wrote:I was doing production before and switched to recognition at around the 1k mark in Core, but I'm still not sure about it. While manga and article reading have definitely improved, I feel I just don't know the words as well. It feels sorta hollow to be able to read it and pronounce it, but not be able to write it if someone asks you. Not actively producing the kanji when SRSing basically means you won't know how to write the word.

Anyone care to humour/question me? Offer more insight? Anyone feel the same way or　go about nihongo with this line of thinking?

*this probably won't be an issue with concrete nouns (e.g. tree, cloud), but there is an issue of idiomatic usage. The expression "use your head" works the same way in Japanese as it does in English, but there are many more expressions that don't have a literal equivallent, and some that even translate into other unrelated expressions (like 足を引っ張る --

)

Animosophy, my personal experience and thinking confirms everything you say.

Animosophy Wrote:It's noted on wikipedia that "it is possible for the productive vocabulary to be larger than the receptive vocabulary, for example in a second-language learner who has learned words through study rather than exposure, and can produce them, but has difficulty recognizing them in conversation", which is fine, I suppose they have the same value.

Thinking hard, this situation is possible indeed, but it's the schoolish way of learning and it's extremely artificial. Having active vocab > passive vocab necessarily means that the active vocab amounts to a few hundred words at most, which is absolutely nothing. It means this learner is at a very elementary level and has NEVER been confronted to native material. If he had, he would have faced hundred of new words in just a few hours : so either he knows them and passive >> active, either he just doesn't get anything. This situation is too exceptional and too limited in time (the first few weeks of learning, or months if you're slow...) to be of any interest.

By the way, 50k passive vocab for a 12 years old? Holy shit, I thought this would be the passive vocab of an educated adult. I'd like to see more references and studies on this later.

Animosophy Wrote:However, it leaves the learner in an awkward position when wanting to use productive vocabulary in contexts they have yet to see it used in, or even worse, associate incorrect contextual uses with a word* (not to mention lacking recognition). Is production vocabulary not better (if not naturally) acquired AFTER significant "receptive control"?

Of course. It's ridiculous to try and use a word/expression that you have learned contextless (the only way I can think about here is from vocabulary lists). Language is 90% about (1) data input (vocab, and kanjis in Japanese) and (2) mimicking sentence structures and idiomatic usage/expressions. The 10% left is grammar, and how smart you are to use it finely. But you can be very smart and read a ton about grammar, know all the words from lists, your Japanese will still be awful if you haven't had sufficient exposure to native Japanese (the (2)).
Also I think the word "significant" in your last sentence is key here. I believe you should absorb as much native material, and for a very long time (or: large quantities in a shorter amount of time, if you can... but we're still talking months here), before doing active effort to improve your production... I'm absolutely not saying to not speak/write at all here. But concretely : finish Core 6k, don't worry about production. When you're done, you'll be able to produce a little, though slowly and in a very limited way, relatively to all the data your absorbed with the 6,000 sentences. But then if you start using Japanese actively (IRL discussion, messaging, mailing, forums), you production ability will quickly and smoothly improve, and it will most probably be "natural" Japanese, because you will have seen enough native Japanese (the sentences) to recognize good and bad patterns. You'll be able to act as your own master, preventing yourself to say incorrect Japanese, because you'll know that it shouldn't be formulated that way, or with these words.

Animosophy Wrote:So when is a good time for production-based learning? I'm going to be hasty and say there is little-to-no place for vocabulary production cards in an SRS environment.

Unless you need/want to improve your production ability (for practical use) before reaching excellent passive understanding (98%+ in both reading and listening), yes, don't even bother with production. I focused fully on reading sentences in Anki (continuing adding sentences of the same format as Core 10k after finishing it), currently I'm at 14.4k sentences/16.4k vocab and I have seen my production ability improve nicely, without ever forcing myself. Then again, I never needed to speak or write Japanese, I just wanted to consume a lot of native material. As time passed I found myself more and more involved with Japanese people and had to use it actively (and of course this made my production ability improve even further), but I wasn't really looking for that. It just comes naturally as your Japanese reaches a "usable" stage (i.e. able to have "smooth" and "not boring" conversations; very far from "fluent" here).
If you have other goals or obligations, the story is entirely different and you will want to work with a different passive/active ratio.

Animosophy Wrote:SThis is from another thread, which I can relate to as a mental barrier preventing me from trusting recognition cards by themselves, without the reassuring expectation of production ability (emphasis mine):

devilsbabe Wrote:I was doing production before and switched to recognition at around the 1k mark in Core, but I'm still not sure about it. While manga and article reading have definitely improved, I feel I just don't know the words as well. It feels sorta hollow to be able to read it and pronounce it, but not be able to write it if someone asks you. Not actively producing the kanji when SRSing basically means you won't know how to write the word.

It's not the same thing here. devilsbabe was not talking about producing natural sentences, word choice and good grammar usage, but only about being able to physically write in Japanese (kana and kanjis). By the way I'm completely unable to do the latter (I think there's a dozen hiragana that I can't even write by hand), and... I couldn't care less, really. It would become a problem if I ever had to teach (or follow a course maybe) in Japanese, but I doubt this will happen anytime soon.


TLDR; the default route is "ingest thousands of words with SRS until you can read and listen to native material; then ingest some more thousands until passive fluency". Now, according to how soon before passive fluency you want to produce Japanese, twist this route by injecting active production situations (artificial with SRS/lang-8 or real with Japanese people). If you are in no hurry for production, the most efficient route is to not study it actively at all I think, because passive feeds active, even if it's at a low rate and in an uncontrolled way.

Phases in acquiring language skills:

PERCEPTION: partial – full
RECOGNITION: partial – full
REPRODUCTION: partial – full
PRODUCTION: partial – full

Even longer read:
http://users.bestweb.net/~siom/martian_m...ssages.htm

[RawToast]

Quote:By the way, 50k passive vocab for a 12 years old? Holy shit, I thought this would be the passive vocab of an educated adult. I'd like to see more references and studies on this later.

I find that strange as well, seeing as wikipedia sates:

Quote:An 1995 study shows that junior-high students would be able to recognize the meanings of about 10,000-12,000 words, while for college students this number grows up to about 12,000-17,000 and for elderly adults up to about 17,000-21,000 or more

I presume how they differentiate what is a 'word' is more lenient for the 50,000 figure.

RawToast Wrote:

Quote:By the way, 50k passive vocab for a 12 years old? Holy shit, I thought this would be the passive vocab of an educated adult. I'd like to see more references and studies on this later.

I find that strange as well, seeing as wikipedia sates:

Quote:An 1995 study shows that junior-high students would be able to recognize the meanings of about 10,000-12,000 words, while for college students this number grows up to about 12,000-17,000 and for elderly adults up to about 17,000-21,000 or more

I presume how they differentiate what is a 'word' is more lenient for the 50,000 figure.

Now that's clearly too low, considering the figures I gave in my previous post...
I'd roughly put a high-schooler at around 25k-30k (possibly 40k) and a college student/educated adult between 30k-60k... That's just an educated guess obtained by crossing my own word count with how far I am from an "educated adult".

Thanks, Warp. You've helped me gather the confidence to add some drastic hourage on my Japanese. The added insight's also appreciated

buonaparte Wrote:Phases in acquiring language skills:

PERCEPTION: partial – full
RECOGNITION: partial – full
REPRODUCTION: partial – full
PRODUCTION: partial – full

Even longer read:
http://users.bestweb.net/~siom/martian_m...ssages.htm

I remember reading that page right before registering here. It had a big impact on my confidence, and I more than appreciate buonaparte's efforts to collect such an amazing pool of resources. I'm spending many hours on the L-R core files every day, and combined with shadowing I'm confident I can bank several hundred hours of speaking practice as well as active listening-reading in a relatively short period of time. Wish I could've let him know and thanked him

Warp2243 Wrote:Now that's clearly too low, considering the figures I gave in my previous post... I'd roughly put a high-schooler at around 25k-30k (possibly 40k) and a college student/educated adult between 30k-60k... That's just an educated guess obtained by crossing my own word count with how far I am from an "educated adult".

If we're talking word families, it should be a bit lower.

Here are the most recent results from an online survey that has been running for a couple of years now and has been completed more than 2 million times.

http://testyourvocab.com/blog/2013-05-10...of-results

And here's a graph showing the vocabulary of native speakers.

http://testyourvocab.com/blog/2013-05-08...ter-detail

Going by the results, very few native speakers have a vocabulary of over 40000 words in English. The people behind the survey have also said that their numbers are above average when compared to SAT scores, so the numbers aren't fully accurate. In fact, they're probably too high.

To me, 12000 for the average middle schooler and somewhere just below 20000 for the average high schooler sounds about right.

The whole anti-production attitude around here is hardly new. As far as I've seen, I'm the only person I know of here who actually attempted and completed core6k in a production style. That being said, I think it was a good experience. Two things that I liked about it are:

1. It got me thinking more critically about definitions as I tried to distinguish between all the synonyms. Its a real challenge. More strenuous in an ACTIVE way and therefore often more memorable.
2. It felt good to be able to actually test WRITING. I have maintained a close bond between the keywords and kanji because of this. I find it actually helps me. The kanji keywords are like part of a new mneumonic that I subconsciously took in each time I was learning a new word.

So my experience has been quite positive, and I don't like the suggestion in this thread that production has no place in language learning. Regardless of which method you think gives the best results, consider the following: BOTH of the above methods result in an APPROXIMATION to a complete understanding of a word.
See, there are 3 things to know: reading, meaning, kanji. There's also three ways of testing them:

production: meaning -> kanji, reading
recognition: kanji - > reading, meaning
audio: reading - > meaning, kanji


A complete understanding of the word means being able to do all three. Using ONE style of card guarantees you will be able to do ONE and NOT necessarily the others. So ALL of the above are just approximations. Each have their strengths and weaknesses. Production stresses writing ability and ability to differentiate between synonyms. It is good for people who want to start speaking/writing right away. Recognition essentially leaves writing untested while stressing interpretation. It is good for people who just only want to read and read and read right off the bat.
So knowing that each has its strengths and weaknesses, why do you want to just arbitrarily eliminate one? Can you not imagine an SRS strategy that makes use of all three?

I don't think I'll ever understand this point of view. I think it stems from the fact that people here think writing down an accurate definition of a Japanese word is some form of mysticism, not to be attempted by mere debutants, rendering production impossible. Its something I can't sympathize with (and before you say it, yes I consult a J-J dictionary and example sentences). I say, the sooner you start thinking critically about definitions, the better. And production essentially forces you to do so. With recognition its easier to gloss over the definition part. And I mean, even if you didn't manage to convey 100% of the meaning of the word with your definition, just like with recognition, your knowledge of a word will be refined with exposure in EXACTLY the same way. You can update it to fit any new earth-shattering revelations, so what's the big deal?

Another thing in this thread is the whole idea that you might screw up if you try to use words you learned from your production cards. All I have to say to that is: WELL DUH. Obviously you could screw up if you're testing out the usage of some new words you learned. That's a good thing! It's why we have things like lang-8. It's part of the process of learning, and putting effort into the production of sentences strikes me as something that is ALWAYS a good idea. Regardless of your level. And as I said before, if you want to do production like this right away, study production cards. It only makes sense.

In the end I think it comes down to a preference. I'm currently using both. If you really don't like production cards, don't use them. But as for tring to prove that you can't derive an equal or greater benefit from them, I think that's just a little bit ridiculous.

Haych, maybe that will surprise you, but I agree with all that you said. You clearly have a lot of experience with the production approach (well no wonder if you finished Core 6k that way...) and what you say is very sensible to me. If anything, it is obvious that doing production cards will result sooner in a refined knowledge of the words.

However what would make me recommend to a beginner doing Core with reading cards instead of production cards is simply an issue of time and quantity. I would be quite interested to know in how much time you completed Core 6k with production style (and, in case you honestly think you were not regular enough or could have been faster (but let's be reasonable here), in how much time a fast learner would have made it. Fast learner with 2-4 hours/day, not hikikomori).

From what I've seen for myself, a few friends and several members reporting their numbers on this forum, fast-paced learners using Core in reading mode learn about 1k cards/month. So it will all come down to the following possibilities :
- Core 10k, reading cards only (no production at all) in 10 months/1 year
- Core 10k, production cards only, in ? (you tell me)
- Core 10k with mixed cards, e.g. you put 40% of the most useful cards as production cards. (then we need a good way of finding those 40% without wasting too much time but I'm sure this is doable). Then in terms of necessary time you will obviously get 40%x(time for production cards) + 60%x(time for reading cards).
- Core 10k with BOTH production and reading cards (and even audio would be great, but I don't want to talk about this here). Of course you'd end up with a great Japanese doing this, but again, time... 1.5x? double, triple the time? Considering we're playing with a base time of 1 year, we don't want to get multiples of this thoughtlessly.

For what it's worth, Haych, I'm completely open to the possibility that what you did is more efficient than what I did, in the following sense : in a 2-year time frame post-RtK, you get a better Japanese with production cards than reading cards. "Better" would be a global score measuring both quantity and quality. And the whole point here is that you get better quality with productions cards but less quantity, and vice-versa, and I think we need to quantify this.
Could you upload (a sample of) your deck please? I would like to test it. The most difficult cards if possible, I don't really want to test it with これは[book]です。

By the way, I would be VERY interested to hear about someone who did Core 6k (or even 10k) with full audio cards, in the following format : front = audio sentence, back = kanji+furigana+english translation, evaluation mode = understanding all words, the grammar, the sentence structure and the global meaning (just like reading).
It's too late for me to do this for Japanese (or maybe not, but I would have to select the 2 or 3k cards that I wouldn't recognize well/not at all if spoken, and finding those will be a bitch). But I intend to do this for Chinese and it will probably be a very interesting experience, I want to measure how much better you can internalize words with SRS'ed audio and at what level of understanding 5k cards would put you when trying to listen to some drama/movie/etc. I have a feeling that there's a highly underused learning method here (at least on this forum I've almost NEVER seen anyone mention massive audio cards, except you in your previous post).

[RawToast]

Haych Wrote:The whole anti-production attitude around here is hardly new. As far as I've seen, I'm the only person I know of here who actually attempted and completed core6k in a production style. That being said, I think it was a good experience.

Yes, this forum does tend to be more in favour of recognition based learning and reading/listening. Both methods are valid, even the usual internet polyglots are divided on this (e.g. Benny and Steve Kaufmann.)

I stick to recognition on Anki, but I do some production on iknow. I do find my retention is slightly better on the iknow words, but that could be down to the increased exposure.

Haych Wrote:I say, the sooner you start thinking critically about definitions, the better. And production essentially forces you to do so.

I hadn't considered this point before, and I agree completely. However, Warp addresses the point on time and quantity which essentially echoes my point of view.

Production-based study requires more mental energy, so it can't be studied at the same volume or pace as recognition, and failure to recognise outside of an SRS has no penalties. It's not like we can actively absorb sensory input all day, but output does have its limitations on this front. It detracts nothing from the value of production itself, only the limits of its application relative to recognition.

Warp2243 Wrote:However what would make me recommend to a beginner doing Core with reading cards instead of production cards is simply an issue of time and quantity.
[...]
From what I've seen for myself, a few friends and several members reporting their numbers on this forum, fast-paced learners using Core in reading mode learn about 1k cards/month. So it will all come down to the following possibilities: [...]

By the way, I would be VERY interested to hear about someone who did Core 6k (or even 10k) with full audio cards, in the following format...

This is exactly what and how I think.

I suppose I'm doing something similar. My medium term goal as of now is to complete Core6k by the end of the year in this way...

- Shadowing and Listening-Reading the core sentences in 5 progressive stages (audio+kana [I may skip this step after core2k] > audio+eng > audio+eng+small kanji > audio+kanji+small english > audio+kanji+sentence analysis)
- Study vocabulary cards through mnemonics in this format -- front: kanji / back: kana, audio and meaning
- Full note-taking of Basic and Intermediate Dictionaries of Japanese Grammar, using uisukii's non-CD DoJG decks in tandem

Ought to give me 400+ hours of speaking, 700+ hours of listening-reading, and a good foothold to familiarise myself with grammar and sentence patterns. I also intend to finish core10k in this way in August next year, just before I leave to study in Japan.

Warp2243 Wrote:However what would make me recommend to a beginner doing Core with reading cards instead of production cards is simply an issue of time and quantity. I would be quite interested to know in how much time you completed Core 6k with production style (and, in case you honestly think you were not regular enough or could have been faster (but let's be reasonable here), in how much time a fast learner would have made it. Fast learner with 2-4 hours/day, not hikikomori).

Looking at some stats, the earliest cards come from 30 weeks ago, but I was struggling with RTK3 reviews and only really picked up the pace 25 weeks ago. So, December is when I started, basically. Finished two weeks ago, so that's about 6-7 months (that's basically 1k/month, so I guess I'm in the green). Anki says average time per day is 91 minutes, but that probably doesn't account for editing time very well, so I'd say I averaged 2 hours a day. Total time on the deck is 314 hours. Per new card added, I think that time might be a bit higher than others I've seen. Normal number of reviews for me were 200 at slow times, and up to 400 at high. Normal numbers for a 5 minute session is 25-30 cards (initially, I was slower because I wanted to write the words, but then I realized I can just call up the keywords for the kanji and leave the test of the writing to my RTK cards). Probably the worst stat is correct %, which usually was 70-80% before adding new cards. My mature % for all time is 83.4%.

So you can guess from that percentage that it wasn't a walk in the park. As for how easy it would be for the average learner, I am not sure. It took up a lot of my time, and I am still a student. For someone who is working full time, it might be less practical. I agree that with this and RTK, people should be aiming to complete it fast. That's why I'd probably only recommend it to someone who has the motivation and time to make it work.

Warp2243 Wrote:- Core 10k with BOTH production and reading cards (and even audio would be great, but I don't want to talk about this here). Of course you'd end up with a great Japanese doing this, but again, time... 1.5x? double, triple the time? Considering we're playing with a base time of 1 year, we don't want to get multiples of this thoughtlessly.

That's an interesting thought. I know the actual smart.fm site uses that. I sort of doubt you would end up doubling your benefit. I also doubt you would double your time. If you know it one way, doing the card going the other way is going to be much quicker. If you're reading native media and you did production cards, 99% of the time you at least recognize if you see something you've studied, but sometimes the definition doesn't come back to you right away. I think the more difficult words might benefit from it, though. I am using some recognition cards with verbs I've studied to try to recall all the various meanings they can take.

Warp2243 Wrote:Could you upload (a sample of) your deck please? I would like to test it. The most difficult cards if possible, I don't really want to test it with これは[book]です。

I don't know a simple way to export sections of a deck, so here is the deck minus scheduling. https://dl.dropboxusercontent.com/u/2771...chC6K.apkg
It is just the 'core6k optimized' deck, so the sentences you see will be stock. All I've done with it is just rewrite definitions that I thought were odd or confusing, and added synonyms and other information to help me out in the 'caution' field, which appears on the front. I don't know if you consider that cheating, but I tried in the beginning, and while there might be minor differences in usage, I found sometimes you can't make a definition different enough. In that case, I need to add it to the front or else I'll just find myself listing off possibilities rather than giving an answer. I also added some other stuff to help me out sometimes, like how many kanji, and if it is onomatopoeia, or verb stems, or whatever. I can get pretty lax when it comes to hints with this deck because it gets tough.

Also, the synonyms I list are only ones from within core6k, so if you know more words, it might be even harder. In my studies outside core (using JLPT1 terms from coreplus in recognition, not core10k), I have come across yet more synonyms that might have to be added to the front. Gotta admit, sometimes I am glad coreplus is recognition because the lists for some definitions like "situation" just started getting ridiculous.

Warp2243 Wrote:By the way, I would be VERY interested to hear about someone who did Core 6k (or even 10k) with full audio cards, in the following format : front = audio sentence, back = kanji+furigana+english translation, evaluation mode = understanding all words, the grammar, the sentence structure and the global meaning (just like reading).

I actually did this for a while, but not with core6k (before my nice 2 year break, not a good sign). It was far too easy to recognize the sentence itself rather than force yourself to freshly interpret it each time (especially if you draw from different sources). Maybe if you did a lot of them, or just did individual words, they would start to blend together and would end up being better for you. It's an interesting thought.

Haych Wrote:The whole anti-production attitude around here is hardly new. As far as I've seen, I'm the only person I know of here who actually attempted and completed core6k in a production style.

I only ever SRSd Japanese in a production style. I of course wasn't producing Japanese text from English text (that would be unnecessarily difficult, and would force you to memorize the specific Japanese words and sentences your material happens to use, rather than just gain the ability to correctly translate English into Japanese).

But producing Kanji from Kana is in fact production. So is producing readings from Kanji. Furthermore, doing both amounts to fully producing Japanese vocab (not grammar though), without the overhead of trying to remember the arbitrary phrasing of a specific answer.

In conclusion, as far as vocab is concerned, the only SRS method that doesn't involve production, in Japanese, would be to provide both the Kanji and the reading of Japanese text in the question, and expect the English translation in the answer. I am not aware of a single person on here who recommends or does that.

As far as grammar, learning to correctly compose sentences in a language, through production, is impossible: it's impossible to reproduce correct Japanese with any reasonable degree of accuracy, from learning a set of rules ; it's also impossible to memorize how to produce all the Japanese sentences one will need to speak Japanese correctly. The only possible way to learn to correctly compose sentences in a language is through listening and reading massive amounts of materials. You can't memorize a language. You can only memorize its words (and even there, only a portion, hence the concept of an active and a passive vocabulary).

@Stansfield, I wholeheartedly agree.

Having kanji and kana on the front of a card is purely recognition. I have the kanji on the front and kana, audio and English at the back.

I'll admit this is not purely recognition, but it works better. Thanks to RTK serving as a kind of semantic reference, I derive and memorise meaning from new words too quickly/easily, even if I make a mistake based on the keywords I've associated to the kanji. In situations like this, I think learning the readings is a logical step. Although, I would argue that having kanji on the front side of a card already involves recognition whether we like it or not; if you can't recognise the meaning, it would be appropriate to fail the card even if you manage to recall the reading. That would be a weird situation for me since I rely on mnemonics to memorise the readings. I'm on my phone, this is the best I could do to make a point

Reading this thread leaves me with a question I would like answered:

What should go on the front, and back, of cards for learning vocabulary?
(assuming the learner has completed RTK1, but knows zero Japanese vocabulary)

English to kana and kanji?

[RawToast]

Doctorate Wrote:Reading this thread leaves me with a question I would like answered:

What should go on the front, and back, of cards for learning vocabulary?
(assuming the learner has completed RTK1, but knows zero Japanese vocabulary)

English to kana and kanji?

Rather than derail this thread, I suggest anyone answering this does so here:
http://forum.koohii.com/showthread.php?tid=10900

Quote:In conclusion, as far as vocab is concerned, the only SRS method that doesn't involve production, in Japanese, would be to provide both the Kanji and the reading of Japanese text in the question, and expect the English translation in the answer. I am not aware of a single person on here who recommends or does that.

That may seem strange in the long term, but I could see it working for the first reviews. You could have those 'pure recognition' having a faster time scale than the usual pseudo-recognition cards. Those cards would fit the 'only learn one thing per card' approach of supermemo.