I do 25 kanji a day. Yesterday I couldn't open Anki and review once. I'm behind, and I can't do new cards today. 50 is overkill. What should I do?!

It's okay to take breaks, this is not a race.

(2016-06-03, 2:51 pm)EratiK Wrote: It's okay to take breaks, this is not a race.

I just looked at Anki and I only have 25 new cards today, which is good. I was worried it'd bump it to 50 for some reason.

If you can't do it all, just do some.  It's important to keep in the habit of doing it but don't overload yourself.

[Zgarbas]

I for one embraced my imminent transformation into a sloth

You can always temporarily change Anki's setting for how many new cards it offers you, so you can work on any built up backlog of "to review" cards without adding to your pile with more "new" ones.

I would advise against using Anki's new cards feature at all.

Standard practice is to suspend all the cards, and selectively unsuspend the ones you want to learn on an ad hoc basis.

Thing is that it's very often the case that adding related cards together makes things a whole lot easier, even if it means adding them in bigger batches (but less frequent).

Edit: I thunk of another thing.

Use reviews for reviewing not learning.

Study the new cards in the browser or in the book first before reviewing them. Look for patterns/differences between related cards that you can use as an aide memoire, informally test yourself by covering up the answers etc, and only then review them in Anki. A bit of up-front investment of time will repay 10x in the subsequent reviews.

(2016-06-03, 4:08 pm)Uanotherjohn Wrote: I would advise against using Anki's new cards feature at all.

Standard practice is to suspend all the cards, and selectively unsuspend the ones you want to learn on an ad hoc basis.

Thing is that it's very often the case that adding related cards together makes things a whole lot easier, even if it means adding them in bigger batches (but less frequent).

Edit: I thunk of another thing.

Use reviews for reviewing not learning.

Study the new cards in the browser or in the book first before reviewing them. Look for patterns/differences between related cards that you can use as an aide memoire, informally test yourself by covering up the answers etc, and only then review them in Anki. A bit of up-front investment of time will repay 10x in the subsequent reviews.

How do I do this on mobile?

I miss reviews all the time. Then again, I am not the best example xD

(2016-06-03, 3:39 pm)Zgarbas Wrote: I for one welcomed my imminent transformation into a sloth

While not adding cards obviously slows down your progress, it's not a big deal to take breaks from adding cards every once in a while. That being said, I highly recommend that you not skip any reviews, no matter what. Unless your computer/cellphone's exploded, which hopefully won't ever happen to you.

(2016-06-03, 4:38 pm)Queen of Mars Wrote:

(2016-06-03, 4:08 pm)Uanotherjohn Wrote: I would advise against using Anki's new cards feature at all.

Standard practice is to suspend all the cards, and selectively unsuspend the ones you want to learn on an ad hoc basis.

Thing is that it's very often the case that adding related cards together makes things a whole lot easier, even if it means adding them in bigger batches (but less frequent).

Edit: I thunk of another thing.

Use reviews for reviewing not learning.

Study the new cards in the browser or in the book first before reviewing them. Look for patterns/differences between related cards that you can use as an aide memoire, informally test yourself by covering up the answers etc, and only then review them in Anki. A bit of up-front investment of time will repay 10x in the subsequent reviews.

How do I do this on mobile?

You can suspend/unsuspend cards one at a time using the card browser on Anki mobile (for android). Just hold your finger on the card until a menu pops up.

However, to mass suspend or unsuspend cards, you need to go into the desktop application.

In the vein of anotherjohn, I'm a big fan of the "Anki is for remembering, not for learning" mindset.

My own approach with my vocab deck is to not add new cards by default. I steal time for reviews during the day using the mobile app. After finishing my reviews, as time allows, I use the Study Ahead feature of the mobile app to do new cards with vocab drawn from my current reading. I take time to review the cards I've added, write them words down, and look up J definitions and an example sentence from Yourei.

I add only as many new cards as I think I can handle that day. If my success rate falls below 90%, I forget new cards entirely and focus on remembering what I've forgotten. Or I suspend/delete the stupid cards if I don't think I'm encountering them enough in the wild (see ya, 跳梁跋扈).

Do like five cards. You're only obligation is five cards. You can do more, but five cards is your obligation.

Congratulations you've finished all 50 cards 20 minutes later. Trust me.

(2016-06-08, 12:13 am)NinKenDo Wrote: Do like five cards. You're only obligation is five cards. You can do more, but five cards is your obligation.

Congratulations you've finished all 50 cards 20 minutes later. Trust me.

Works the same way for adding kanji.  I set my obligation at just one character each day, but I don't think I ever actually did less than 10.

"Oh, the next character is just the same thing with one primitive added!"
"Oh, I'll just check if anyone has a really good story for the next character."
"Well, there's just one character left until I'm done with this primitive, so I might as well finish it off...."

Random question but should I be recognition from the kanji, the keyword, or both? Because my deck asks both. It'll quiz on what the keyword is, showing the kanji and it'll also show the keyword and you have to write the kanji. Which is better?

(2016-06-11, 10:14 pm)Mars Prism Makeup Wrote: Random question but should I be recognition from the kanji, the keyword, or both? Because my deck asks both. It'll quiz on what the keyword is, showing the kanji and it'll also show the keyword and you have to write the kanji. Which is better?

If you want to write them, do keyword->kanji, if you're just doing RTK to get better at recognising kanji, do kanji->keyword. I see no reason to do both unless you really want to.

Mars Prism Makeup Wrote:Random question but should I be recognition from the kanji, the keyword, or both? Because my deck asks both. It'll quiz on what the keyword is, showing the kanji and it'll also show the keyword and you have to write the kanji. Which is better?

I've always done keyword->kanji and required correctly writing the character to pass.

Honestly I don't see any reason to do kanji->keyword, if you want to drill on pure recognition then why bother with RTK, and why not just skip straight into brute-forcing vocab words with kanji?

Still, many people -do- drill kanji to keyword and say it helps, which makes no sense to me... but then again learning styles differ. I've always found a strong connection between writing out my study materials and retention of what I'm learning about... in college I became a big note-taker both in class and while reading my texts for just that reason. It actually turns out that taking handwritten notes is -much- better for retention than typing notes, and any kind of note-taking is better than just recording the class for a later re-listen. Of course writing out reviews is not quite the same as note-taking, but it has similarities in the time taken where you're forced to think about the material and the engagement of multiple senses in the learning process.


For what it's worth, I'd say you in particular should be drilling keyword->kanji, since you have fluent friends and intend to communicate in Japanese, I can easily imagine that you'll be writing postcards or something where you do actually need to be able to write the kanji. If I'm wrong about that and you don't care about being able to write, then I suppose there's merit to looking at doing it the other way around. There are reasons to think that drilling kanji->keyword gets you to reading faster (which is presumably your primary goal), at the cost of not being able to write.

Edit: Also... I wanted to ask, how did you get dethroned? You used to be a Queen!

There is no reason not to drill keyword to Kanji. Heisig is actually BAD if you do it Kanji to Keyword, as you're settinhg up a really shitty mental association that will be way harder to get rid of than the other way around.

The keywords are meant to be prompts for recalling the production of the kanji, and as association anchors for when you start sticking them in vocabulary. You are not really supposed to think of the keyword when you SEE the kanji, not as an automatic thing anyway. Even at the height of RTK, I would have to think quite hard upon seeing a kanji to conjure up its keyword, and that's exactly what you want.

Why?

Because KEYWORDS AREN'T THE MEANING OF THE KANJI!

Sometimes they have little to nothing to do with ANY meaning of the kanji, often they have a connection only to one limited use of the kanji, and as you get further in, and Heisig starts using synonyms in English, the difference between the nuance of these synonyms often has absolutely nothing to do with the nuance between the two kanji bound to these synonyms.

The keywords are, at best, meant to hang on long enough to use them a mnemonic anchors as you start learning vocabulary. All but the most simple ones (the ones that essentially DO boil down to the meaning of the kanji) should disappear fairly quickly.

Doing it both ways is not simply inefficient, it's detrimental.

(2016-06-11, 11:44 pm)NinKenDo Wrote: There is no reason not to drill keyword to Kanji. Heisig is actually BAD if you do it Kanji to Keyword, as you're settinhg up a really shitty mental association that will be way harder to get rid of than the other way around.

I don't really disagree with you, but I do think you're overstating the case.
One of the massively important things that Heisig does for you is give you the ability to recognize kanji as individual entities. Hanging a unique mental tag on them in your native language is a huge part of that, though it's usually overlooked.
That part of the system should still work even going Kanji->keyword.
Of course, the recognition of individual components and the ability to correctly interpret stroke order and so on will only come with keyword->kanji (along with actually or mentally writing your answers).


Also of course, some people -do- have trouble recognizing kanji that they've learned with keyword->kanji practice. I've never really had that problem or never seen it as a problem. Sure, a few times I've only realized which keyword a 'known' kanji is after looking up a word in the dictionary, but so what? Having once made that association, I never have that problem again for that particular character. Still, learning styles differ, and for someone who really has trouble reversing the associations when they start reading Japanese I can see why reverse-reviews might make sense. I don't think that the Average Learner™ will have that problem though.

(2016-06-12, 12:05 am)SomeCallMeChris Wrote: One of the massively important things that Heisig does for you is give you the ability to recognize kanji as individual entities.

Is that something many people have trouble with in the traditional non-Heisig approach to kanji? I don't recall it ever being much of a problem for me: kanji weren't introduced too many at a time and they're associated with vocabulary you're learning so that's the mental tag you associate them with. The hard part was being able to correctly write them, which is not something you're going to be able to fix if you only review kanji-to-keyword.

Conversely I only ever did Heisig reviews as keyword->kanji and so I can't remember the keyword given the kanji at all.

RTK is about learning an English reading (what Heisig calls an 英読み) for each kanji, putting non-漢字圏 learners on the same footing as e.g. Chinese learners, who start with a massive advantage despite knowing lots of deeply-ingrained 'wrong' things about Japanese kanji.

That is the point.

SomeCallMeChris Wrote:Of course, the recognition of individual components and the ability to correctly interpret stroke order and so on will only come with keyword->kanji (along with actually or mentally writing your answers).

Recognising the kanji components is an absolutely crucial part of associating the kanji with its meaning. It's pretty well impossible to do it without.

And I don't understand your point about stroke order with regard to reading.

And to learn the reverse association through random exposure is to forget the point of SRS.

When I switched to reviewing kanji -> meaning my accuracy dropped from the high 90s to about 80%, much of the drop caused by mixing up similar-looking kanji or failing to recognise components due "mentally writing" them and thus hardly ever having actually looked at the ruddy things during reviews. This I found somewhat irksome after having been given so many confident assurances to the contrary. I guess I must not be up to the standard of the Average Learner™.

Based on this *experience* it is my view that reviewing keyword -> kanji wasted an awful lot of time. This may have been mitigated by actually writing them, but that would have taken even longer, and anyway learning to write without being able to read creates artificial problems (similar keywords, positioning of elements) entirely due to the absurdity of it.

If/when I finally do get around to learning handwriting (which I have no use for but kinda sorta want to do it anyway), I will be in a position to do it entirely in Japanese, using Japanese radical names / mnemonics / etc - I prospect I find quite appealing.

(2016-06-12, 12:05 am)SomeCallMeChris Wrote: Also of course, some people -do- have trouble recognizing kanji that they've learned with keyword->kanji practice. I've never really had that problem or never seen it as a problem. Sure, a few times I've only realized which keyword a 'known' kanji is after looking up a word in the dictionary, but so what? Having once made that association, I never have that problem again for that particular character. Still, learning styles differ, and for someone who really has trouble reversing the associations when they start reading Japanese I can see why reverse-reviews might make sense. I don't think that the Average Learner™ will have that problem though.

My problem was exactly the opposite. Before I figured out how to study, I tried learning vocabulary with their kanji (on some learning site or another); I couldn't remember any of them by recognition, because I couldn't distinguish kanji very well.

Until recently, I only bothered with the basics of writing kanji (and looked at the stroke order of a few kanji in RTK, Denshi Jisho, etc), but I'm picking them up with little effort. I don't think that would have been true a few years ago, but writing is coming quite easily now.

I've always recommended ignoring writing in the beginning unless you absolutely need it (OP might need it?); it just takes so much extra time. I finished going through RTK in three or four months, if I remember correctly; there's no way I would have been able to do that if I were writing everything.
So I really only used RTK as a springboard into vocab recognition.

@NinKenDo
The keywords start fading almost immediately after you start learning vocabulary. For an example (anecdotal), when I was studying kanji recognition with English keywords, I saw 火山 and thought "fire mountain, so 'volcano'", I see those same kanji now, and those keywords aren't the first thing I think; it's now mixed with everything else I know about those kanji and that word, but the word itself is front and center, so more like "that's かざん, the geological feature that spits out molten rock, with the kanji 火 and 山, 'volcano' in English", and any number of slight associations; even if the keywords are obvious, they are mute compared to all the other things.

TL;DR
I don't think keywords are any kind of detriment to learning vocabulary, because familiarization with vocabulary and kanji is a great pile of associations that, all together, create the flavor and image of the word.

(This time I was smart enough to pull out my laptop instead of typing walls of text on my phone...)

While some keywords are definitely iffy or totally off, most of them are actually fairly helpful in my experience. Honestly, I really don't get it when people go around saying that you don't learn the meaning of stuff when you do RTK.

For instance, I've reached the 4k mark with Core 6k fairly recently. Lately, I feel several of the new words I learn are at the very least somewhat related to the keywords I learned, if not fairly predictable. In my reviews, there's always a word here and there that makes me feel like I don't really know it... but because I'm used to those kanji readings, I can safely get the pronunciation right, and by looking at the kanji I get a pretty good idea of what the word means. If anything, it's actually annoying when a new word is one of the few cases where my knowledge of kanji doesn't help understand the meaning at all. It's definitely noticeable.

Even if a kanji has more than one meaning which you may not have learned with RTK, it's not like you have to learn all of them right away anyway. I do fail to remember keywords all the time, but it's not just the keywords. Your stories can usually help you remember a little bit of the nuance. Sometimes, you can't remember either, but have a good idea what that kanji means anyway. Another thing that I do sometimes is look up the meaning of the kanji and read my story again briefly (if I feel the need and can't remember) depending on the word I'm learning.

There's also the advantage of being able to look up a kanji based on the keyword at jisho.org if for some reason you feel inclined to do so (e.g. you're reading manga and don't know the reading, the dictionary won't recognize the kanji you draw, you don't want to look it up by radical, etc). It's not that hard to just type the keyword or something similar. Chances are the kanji you're looking for will appear on the right or be displayed in one of the words below.