This:

Input less, output more

(also read the video description)

What do you think of this? How this translate in terms of japanese learning?

After switching my RtK deck a couple of time from "keyword > kanji" to "kanji > keyword" I understand by myself that the first method gives better kanji retention, so although "kanji > keyword" is simpler, "keyword > kanji" is more effective in terms of carving the kanji in long term memory.

Is this considered output?

If yes, what is an equivalent for kanji compounds? For example in sentences, cloze deletion is considered like a sort of output? It will be more effective in long term memory than simple word/sentence recognition? I do think so, what is your experience in regard to this?

Another interesting video by the same person:

I don't think producing kanji is really output. Producing actual Japanese is.

Now this may go against the popular opinion here, but I think going keyword -> kanji doesn't do you any good if you don't care about being able to write. I switched to kanji -> keyword after I finished the book because I was fed up with not being able to recognize a kanji I see in text, even though I've learned it keyword -> kanji.

People keep saying keyword -> kanji makes you remember them better, but I haven't seen any evidence for this yet. Going kanji -> keyword made me actually recognize them when I see them, so from my experience it's the other way round. Sure if you want to be able to write them, you have to go keyword -> kanji. But for reading I don't see the point.

Also just because Heisig did it that way doesn't mean it's the better way to do it. I think the merit of RTK lies in the stories and breaking up the kanji, not in the direction you review them.

Also a huge bonus is reviews are like at least 5 times as fast and you can do them in the bus or tram where you can't write anything because it's so shaky.

I must say I have not difficulties in recognizing the kanji I already know. Maybe the first couple of time I cannot recall the keyword, but it is enough to look for them only those couple of times and from that moment I'll recognize them without difficulties. Also when you read you don't read kanji -> keyword but kanji compound -> meaning of the word, which is different.

For example today I studied 発 (keyword= discharge), but how will this help when I'll encounter words like 開発, 活発, 始発, 自発的, 摘発 or 発育?

Obviously every time I'll see 発 I'll have no difficulty in thinking "discharge", although I go from keyword to kanji, but going this way I feel the kanji is fixed better in my mind. And I'm saying this based on my trials and errors, not because I think Heisig is God, on the contrary in the beginning I was thinking exactly what you said, but I tried and it doesn't worked for me. But this is me, I'm not saying this is the truth or the only way, maybe for me it works better

But I am wondering if it's the same thing for kanji compounds, that is, being forced to recall the kanji compund by memory via cloze deletion is more effective than simply passively recognize a kanji compunds that is already here and all you have to do is remembering what it means? Obviously both ways work, but is the first better for long time retention / more in depth memorization?

Also I do think actual output is japanese production, but obviously we cannot produce it all the time, so beside that, if we all use anki to complement true immersion (both input and output), what is the best way to use it for compounds reviews?

This has been discussed a lot, search for production vs recognition or something like that.

I think both directions are considered "output" in his video, with respect to RTK kanji knowledge (basically a keyword-kanji association). Going kanji -> keyword you are outputting keyword (so you practice identifying the kanji). Going keyword -> kanji you are outputting the shape of the kanji (so you practice retrieving the shape from memory). Both ways are important for kanji knowledge depending on what you want to do.

What he is arguing against is reading the textbook (e.g. biology, Japanese, etc) over and over again in the hope that you will somehow remember it, or thinking that just because you can understand the information well you can actually retrieve the information from memory and use it. You'd be surprised how many people "study" memorization-heavy subjects in university by just reading the textbook and don't use efficient and effective tools like Anki.

To apply to Japanese, I suppose if you just read RTK (the book) 10 times you probably won't remember very much. You have to specifically exercise that memory. To extrapolate from that, it seems to me that if you want to be good at recognition then you should do recognition, and if you want to be good at production you should do production.

Now since production and recognition (in Anki) are just different aspects of outputting the same knowledge (production = outputting word, recognition = outputting meaning), they complement each other well and doing one will to some extent help with the other (but obviously not as effective as doing the other).

There's also the aspect that production does give you a clearer picture in memory - if you know how to write each kanji, it's easy to remember their differences.

He doesn't talk about ANKI-esque things at all. He means you should be able to "say in your own words" and to produce your own stuff after a concept has been explained and examples have been given. Say, for example, your textbook says that you use ながら to express the idea of "doing something while doing something else", and you use it with the stem of a verb etc, and then it gives the example of 食べながらテレビを見る。 He asks you to produce stuff like 歩きながらスマホを見る or even 通勤しながら新聞を読みながらカレーパンを食べる as he thinks that you didn't really understand what you can't produce, and you can't keep in memory what you didn't really understand.
His talk has nothing to do with ANKI-style memorization. Whether you go keyword to Kanji or Kanji to keyword, that's nothing you have to understand, nothing to come up with on your own.
He does speak about memorization in other videos though.

I agree with the idea that you can't remember what you don't understand, but understanding doesn't require physical production (like writing kanji over and over).

It's anecdotal evidence, but I never actively took notes during lectures, except for one class (and that's because we didn't use a textbook and the tests were open note), because, other than the fact that I don't usually find notes that helpful (unless we don't have or use a textbook), what's in my notes doesn't compensate for full attention during the lecture.
If you pay attention to what you're learning and think about it, its implications, and its applications, then you can understand it more easily than if you just try to cram information from your notes or a book.
In a way, this is production, and I think that is what this video is saying (much better than I did). I don't think it has anything to do with study methods used in Anki or anything else.

Honestly, I feel like writing the kanji loses its importance after the first three or so times; which is just to help you remember the basic idea of how to write them. For learning to read with kanji, you only need to be able to easily make entries in your memory for them; that's what I found RTK useful for, being able to 'see' the kanji (before doing RTK, I couldn't really tell one kanji from another, so if they looked even remotely similar, I'd mix them up). After that, though, there's really not much point in RTK, because you'll continue to reinforce and expand your knowledge of kanji by reading.

That's not to say that production practice, in our sense of it, is useless; I can't produce much of anything since I never practice it, but I can understand and read plenty of things; the knowledge is there, I just haven't made the connections to actually use them myself (that's one of my goals for the year, listening is another). I just think that's completely different to what this video was about.

So I've misunderstood the meaning of that video! Thank you all for the clarifications. I still think the act of having to recall the kanji from the keyword, and the compounds from the meaning has a value in in-depth memorization, but maybe the extra time needed it's not worth the effort.

Also I read now a comment of Steve Kauffman about that same video:

"Yes but in his case he is referring to learning facts, like a history text. Reading a text over and over is not as effective as doing an assignment on it. Vocabulary acquisition in languages is different. Each new context is a new learning environment. Therefore we need to move on to new and more challenging material all the time. And of course at some point we need to increase our output, but , in my view, only when we have enough vocabulary and sufficient level of understanding of the language."

It seems coherent to what you all said, so maybe I'll try recognition only for vocabulary/sentences and see if it is enough for me, and I'll spend the extra time reading and maybe producing simple sentences as you suggested

People whose mother language is a Romance one often have a frustrating experience when they try to learn another Romance language: they can understand most of what they read, but the new words never stick no matter how often they come across them and they can't talk worth beans. This happened to me when learning French, because all I did was read books.

If you're learning vocabulary, it's harder for words to stick if you're going Japanese to English. Besides, producing sentences reinforces your knowledge of grammar and of the nuances of the words.

If you're learning kanji the RTK way, keyword to kanji makes things stick more easily. You don't remember words you don't use, you don't remember mathematical formulae you don't use, and you don't remember credit card passwords you don't use. Or at least in my experience it has been like that.

cophnia61 Wrote:This:

Input less, output more

(also read the video description)

What do you think of this? How this translate in terms of japanese learning?

It doesn't. He isn't talking about language learning. You can't do the things he's talking about, in language learning.

The only thing you could do, as a beginner/intermediate language learner is memorize sentences and patterns, through "output". But that's not the kind of output he is talking about. For what he's talking about, it is essential that you master the language you're using already, so that you can focus on producing the content.

It's a good advice for everything except language learning. I use it all the time when "reading" books about programming, for instance. I open up SublimeText, and write, test and debug every single piece of example code that's in it. I spend maybe 10% of my time actually looking at the book, and the rest producing the code in it. And the test of whether I did a good job isn't that it's exactly the same as in the book, it's that it follows the general principles of the book, and, most importantly, that it works. But that's not something you can do with content in Japanese, especially as a self learner. "Output" in Japanese would be the equivalent of memorizing the code, line by line, and then reproducing it without much understanding of the overall structure, and testing yourself by making sure it's identical in every way to your source material.

If you told the speaker in that video that that's how you interpret his advice, he'd likely be horrified. That's a terribly inefficient way to learn. If he has a teacher, I suppose an intermediate learned could spend some of their time trying to produce output, with the teacher hovering over him and correcting and explaining every mistake he makes. But it's still a relatively tedious process, and the explanations aren't as valuable as with other types of learning (because languages are a lot less logical than other materials one tries to learn). So, even in that case, the advice in the video applies only partially.

-

Well what the guy explains in the video is actually what Michel Thomas does, so it can indeed be applied to language learning, but I guess he was talking more about learning in general.