I just remembered that someone a while ago mentioned that Anki is actually only good
to engrave into your brain what you've alerady LEARNED. My question is, is it posibble to learn with anki as well? I mean, that if I just enter some facts / data which I've never heard before on a few cards, and then start reviewing and sooner or later it will sitck (given that I understand those facts), or do I have to learn them from a textbook (for instance), the normal way, and then If I don't want to forget them for the sake of my long term memory I enter them into Anki and keep reviewing till the end

Do you deliberately design your threads to provoke me into rambling? Geez. ;p

My longstanding and I think perhaps minority opinion is that the SRS is the perfect HUD/hub for combining learning + memorizing, conflating information retention and processing, and implementing a spaced multimodal integration system, amongst other purposes. Most of my comments are on that topic, methinks. I call the SRS a 'heterophenomenological life partner' though perhaps metacognitive/metamemory scaffolding is a better way to think of it...

Yeah it is possible to learn with Anki, especially if you use the 'Learn' plugin or 'cram mode' for the initial exposure.

nest0r Wrote:I call the SRS a 'heterophenomenological life partner' though perhaps metacognitive/metamemory scaffolding is a better way to think of it...

Somebody needs to lay off the morning halucinogens!

Depends upon your definition of learn. By my definition, you can not. The SRS pushes stuff into your long term memory, stuff you know (that is, you have learned it). If you don't know it, it's pushing stuff you don't know into your long term memory, that makes no sense, how can you remember something in the long term if you don't even know it?

That doesn't mean you have to use a textbook or learn anything manually. It just means that you can't put 200 unknown kanji in an SRS deck and just review them stupidly. You can put 200 kanji in, and everytime a kanji comes up you can analyze it, check it's meaning and make a story, THEN you can push next. And that's all there is to learning really. If you put stuff you understand in your SRS, you have already learned it.

The "cram" method is pretty good for learning, in a sense. It's a temporary deck for a reason -- so you can utilize it to "learn" the stuff. I don't mean just go through it, hit "easy" or "hard" or whatever.

Say you have a word list. Depending on the importance, I like the Iverson method, but I've been experimenting with this:
Take word list (possibly saved from Rikaichan) and make temporary deck, with both production and recognition cards.
"Cram" these, but each time you come to them, study them -- don't just look at it and be like "I know it, next"
Since "Cram's" intervals are shorter than the actual deck, they will come up more often, especially if you forget them in between intervals.

Once you feel they are good, get rid of the cram deck, and the next day (or later in the real day, if you feel you know/understand them), go through the real deck with the full intervals.

Edit: A friend of mine says that he likes Mnemosyne better because he says it's "a learning program, whereas Anki is a testing program."
What he means by this, I don't really know -- I haven't messed around with Mnemosyne, but I don't think it's any more of a "learning program" than Anki is.

Tobberoth Wrote:how can you remember something in the long term if you don't even know it?

Hmm...being exposed to it between delicatly designed intervals (by anki) I guess should do the trick? Seriously, if you are exposed to sg from time to time, eventually you'll become more familiar with it (recognition), and with a little more time, you should be able to leran it too (production).

Tobberoth Wrote:It just means that you can't put 200 unknown kanji in an SRS deck and just review them stupidly.

I guess you didn't hear about the guy who claimed to have learned all the Heisig kanji within 3 or 4 weeks just by putting them in Anki and reviewing them...or something like that, it was in last September if I remember correctly.

Raschaverak Wrote:the guy who claimed to have learned all the Heisig kanji within 3 or 4 weeks just by putting them in Anki and reviewing them

Doesn't this completely defeat the purpose of RtK in the first place? This just brings it back to rote memorization again. Sure you can "learn" 2042 kanji in 3 weeks, but how is his retention after a while?

It seems like what you're looking for something to take all the work out of learning. Anki has got the long-term covered, which, in my opinion, is a miracle in itself. But getting it in your brain in the first place, I don't think that's something you can passively sit back and let a computer do for you.
I don't have any evidence to support this, but going through the initial learning process is something that will create connections in your brain in ways that SRSing over and over just can't do.

Raschaverak Wrote:

Tobberoth Wrote:how can you remember something in the long term if you don't even know it?

Hmm...being exposed to it between delicatly designed intervals (by anki) I guess should do the trick? Seriously, if you are exposed to sg from time to time, eventually you'll become more familiar with it (recognition), and with a little more time, you should be able to leran it too (production).

Learn and production is not the same by my definition of learn. It's not even the same kind of concept. You can't produce, or even recognize, something you don't know. You can't remember something you don't know. And to know something, you have to learn of it.

If you can't learn with the SRS then what the hell is Incremental Reading... That's silly. Of course you can. It just depends on how you use/structure your cards, and what you set as your task. It also depends on what you are trying to learn. Do you need to memorize a set of dates? That's easy to learn with SRS. Just fail them until you know them. Need to learn a complex philosophical concept? Well, that would probably be easier in a lecture/discussion environment, but also quite possible to set up incremental reading cards that ease you into the concepts, teach you the philosophical vocabulary, etc.

Quote:You can put 200 kanji in, and everytime a kanji comes up you can analyze it, check it's meaning

Isn't this what ANY reasonable person would do with a kanji they didn't know the answer for? I don't know what you mean by "review stupidly"? You mean just clicking the button without doing any thinking about it? Yeah, that would be pretty stupid. But, I assume anyone using an SRS in the first place would actually attempt to answer the question, then take a moment to look at the answer if they failed it...

kendo99 Wrote:

Quote:You can put 200 kanji in, and everytime a kanji comes up you can analyze it, check it's meaning

Isn't this what ANY reasonable person would do with a kanji they didn't know the answer for? I don't know what you mean by "review stupidly"? You mean just clicking the button without doing any thinking about it? Yeah, that would be pretty stupid. But, I assume anyone using an SRS in the first place would actually attempt to answer the question, then take a moment to look at the answer if they failed it...

Yeah, any reasonable person would do it, because they know you don't learn from the SRS, you memorize from it. SRS is just that things are shown to you in repeated intervals. That, by itself, doesn't constitute learning. The learning is what you do before you start pushing the buttons and putting stuff into the SRS.

No, the SRS isn't some magic pill that removes any need to think about what your learning in order to internalize, but it can be used as a platform for both learning (because if you set your cards up well all the information neccessary is there) AND memorization at the same time. Also, kanji is really not the best example, because kanji CAN be learned through rote memorization, it just takes much, much longer if you aren't breaking it into components, using logic to aid understanding, etc.

kendo99 Wrote:No, the SRS isn't some magic pill that removes any need to think about what your learning in order to internalize, but it can be used as a platform for both learning (because if you set your cards up well all the information neccessary is there) AND memorization at the same time. Also, kanji is really not the best example, because kanji CAN be learned through rote memorization, it just takes much, much longer if you aren't breaking it into components, using logic to aid understanding, etc.

But there's no reason to call it a platform for learning. What you do in that initial step in the SRS can be done anywhere. In a book, on a website, in a word document. It really has nothing to do with the actual SRS program, what is important how you learn, not in what program the text you're reading is displayed. (I agree with a previous notion of cramming where the actual software is actually used as a tool to improve your initial learning. On the other hand, the cramming mode in Anki is, at least IMO, not very good. Iverson lists or whatever are probably a lot more effective.)

Ok, a lot of words are being thrown around but nothing really defined.

Someone give a methodological definition for "learn" and "memorized."

Cause I'm seeing a lot of "Anki helps you memorize and retain what you learned." But isn't learning merely initial memorization?

If I put sentences into anki that are completely unfamiliar to me, there is a good chance that it will take that card several months to make it out of the "young card" phase. Often, it will take me a couple of weeks just to get past an interval of 5 days or so.

But if I have a good understanding of the words/sentence, and I can remember the readings longer than a few minutes, then that card will probably progress through the intervals rather well.

vix, I think that depends, sometimes concepts are something you have to "understand" to learn, mathematical concepts for instance. Even if you can spout off all the times tables, if you aren't also aware that 3x3 is also 3+3+3 you don't really "know" multiplication. But, of course that "understood" concept does have to initially move into memory or no learning took place. But there are instance we call learning that ARE just rote memorization. For instance, when we do say someone knows their times tables, or history dates, etc. And Kanji CAN be learned this way, it's just slow and inefficient.

Tobberoth, that's fine and all, but it makes sense to learn through the SRS because once its learned in that specific program, then it's also ready to review for retention. Very convenient, and that's why I think it's a platform for both. Also, if you had looked through the link on Incremental Reading, you'd see that the SRS can be used in a very efficient manner purely for reading material and learning. And it actually works in practice to be a vast improvement over just reading over texts BECAUSE of the SRS format.

I'm going to revise that: It is ALL memorization, it's just some "facts" need a certain amount of logic/comprehension applied before they can be memorized.

kendo99 Wrote:it's just some "facts" need a certain amount of logic/comprehension applied before they can be memorized.

Facts may be encoded in many ways, and some encodings are easier to memorize than others (a kanji may be encoded as an image, as a sequence of strokes, as a list of smaller kanji, etc). It's a good idea to try and find good, easier to memorize encodings before throwing facts into an SRS, otherwise you may end up spending your resources (time, mental stamina) trying to memorize bad, easier to forget encodings.

kendo99 Wrote:It is ALL memorization

Agreed. You can memorize the data, you can memorize an algorithm to generate the data, you can memorize a method to create an algorithm that generates the data; but in the end, it's all memorization.

ps: Just found out wikipedia actually has an article on that.

I think it best to think of the SRS as what it is: customizable software, an interface for information with semi-automatic, user-dependent spacing behind it. Breaking down all the imaginary demarcations and rigidly reactive (apathetic) applications to the program and content, having an abstract enough mindset to be able to flexibly extract (when not prefab), design, and structure information according to the best possible encoding, taking into account the content and how the brain processes it, set and setting (nod to Blahah's hallucinogen reference) of study and use, maintaining dynamic metacognizance amidst this structuring and self-assessment per item, per review, over time, as it becomes progressively, interactively/distributedly internalized via explicit and implicit processes and modalities, transitioning from calculation to retrieval to varying degrees.

Or something. ;p

nest0r Wrote:I think it best to think of the SRS as what it is: customizable software

Before you descended into the eighth chamber of lexicographic madness you made a good point: Anki (and other SRS) aren't a rigid or prescribed tool. They are frameworks, which you can use to do pretty much anything which involves, facts, repetition and rating (actually you don't even have to use rating).

Tobberoth Wrote:Depends upon your definition of learn. By my definition, you can not. The SRS pushes stuff into your long term memory, stuff you know (that is, you have learned it). If you don't know it, it's pushing stuff you don't know into your long term memory, that makes no sense, how can you remember something in the long term if you don't even know it?

To claim it's not possible to learn with such an adaptable piece of kit is absurd, it's like claiming you can't learn with a pen and a piece of paper. Of course you can learn with Anki, in many ways. It really doesn't matter if you want to redefine the word learn, confuse it with memorize or make any other semantic tangents, you can do all of those things with Anki. I think I understand what you're getting at - that you can't effectively learn facts by the usual SRS method - but that's not really what you said.

Tobberoth Wrote:That doesn't mean you have to use a textbook or learn anything manually. It just means that you can't put 200 unknown kanji in an SRS deck and just review them stupidly. You can put 200 kanji in, and everytime a kanji comes up you can analyze it, check it's meaning and make a story, THEN you can push next. And that's all there is to learning really. If you put stuff you understand in your SRS, you have already learned it.

This sounds like learning to me... using your SRS software in this way would be using it to learn.

Blahah Wrote:

Tobberoth Wrote:That doesn't mean you have to use a textbook or learn anything manually. It just means that you can't put 200 unknown kanji in an SRS deck and just review them stupidly. You can put 200 kanji in, and everytime a kanji comes up you can analyze it, check it's meaning and make a story, THEN you can push next. And that's all there is to learning really. If you put stuff you understand in your SRS, you have already learned it.

This sounds like learning to me... using your SRS software in this way would be using it to learn.

IMO, you're not using your SRS software to learn in this case. When you split a kanji into primitives, you use a such a technique to learn. When you make up a visual story, you use such a technique to learn. But you're not using the technique of SRS to LEARN anything, you just happen to see stuff in a certain software, which software it is doesn't impact your learning. It's not important to the actual step of learning.

I know I'm discussing semantics, but this is why I and I believe many others say you should only enter stuff you understand into an SRS, because people might get the odd idea that simply putting a massive amount of stuff you don't know nor understand in an SRS will magically make you learn it. Some people on this forum has for example asked what the point of RtK is when you can simply download a kanji deck on Anki. That's an attempt to learn by SRS instead of by mnemonic tactics. It's like rote memorization, but less effective (since SRS softwares space things too far apart, they depend on understanding or mnemonic techniques, the things you use to learn).

Whether you're going to rote learn with the spacing effect in SRS or encode and retrieve in the SRS, I think that's better than treating SRS as something isolated that you shouldn't study within. In fact it's an argument to the opposite of conflating terms and dismissing 'learning in the SRS' in favour of some reactive, minimalist usage, and instead is a good reason to stress metacognition and a proactive stance towards design and study during SRS sessions, treating it like a HUD of sorts, and integrating it with non-SRS study/immersion/etc.

nest0r Wrote:Whether you're going to rote learn with the spacing effect in SRS or encode and retrieve in the SRS, I think that's better than treating SRS as something isolated that you shouldn't study within. In fact it's an argument to the opposite of conflating terms and dismissing 'learning in the SRS' in favour of some reactive, minimalist usage, and instead is a good reason to stress metacognition and a proactive stance towards design and study during SRS sessions, treating it like a HUD of sorts, and integrating it with non-SRS study/immersion/etc.

It would be easier to reply to your post if you used standard English instead of techno-babble.

Regardless, no one is saying you can't use the SRS window instead of a paper or anything else to learn. Question is why you would have this as a goal, there's really not much to gain from it. It's like people who wonder if they can write an essay in Dreamweaver. Sure, but why? The SRS wasn't made for that.

The Unix principle for the win:
* Make each program do one thing well. To do a new job, build afresh rather than complicate old programs by adding new features.

Not saying that papers were made to learn Japanese, but the idea is more that instead of using the SRS as a "HUD" to learning, make a proper tool for it instead.

With few exceptions, depending on material and strategy: The encoding and retrieval environments should be multimedia, using multiple complementary layers, and should be as similar as possible to one another, preferably identical, and as identical as possible to the usage environment and materials as well. A simple primer that I've referenced before is Medina's book Brain Rules, though pretty much all my references tie in to this idea, as well as the strategy/material-dependent variations.

Once you understand it doesn't matter what you're using, as long as you can implement it in such a customizable way with the spacing effect implemented behind it, subject to 'user' control, then it doesn't really matter how many tools you use for it, though I'd hope if you can do it with one customizable piece of software with a thriving community around it, you logically and parsimoniously would.

Quote:The Unix principle for the win:
* Make each program do one thing well.

This is why unix will never be popular with the general public...unitaskers suck, ask any cook.

kendo99 Wrote:unitaskers suck, ask any cook.

But then you can put them all together, and only use the things you need, rather than have bulky crap that's unneeded/unwanted. Plus, then it's much more re-usable in other, new programs.
Many unitaskers = more useful than one multitasker.