I've decided not to collect and SRS sentences when learning new words. I've just started a new Anki deck which will be a long running one of new vocab I pick up. Will just be single words in isolation (shock horror) and doing recognition from Kanji - Kana/english meaning.

Why? It's about 10x faster than doing entire sentence reviews. I'm sick of Anki eating up all my time that i'd rather be reading real Japanese. This decision comes after my immersion environment really teaching me things like it used to way back when it was the only way I learned before RTK (or in the beginning phases).

Not saying sentences are bad (course not!). One of the cons that really got me about it was learning a word but as part of a sentence and not being able to read it and or recall it's meaning in a different context. That was the exception rather than the rule but I can see how individual words will make me more certain that my reading of the word is correct and I do know it's meaning.

I'm putting in monolingual definitions of words when I can understand the definition 100% monolingually. If not then English is fine. I'm not worried about English interfering with my Japanese because I find the more I learn, the less I have to translate in my head and can just accept meaning for what it is in Japanese.

For me the main barrier to understanding Japanese right now is vocab. 3000 - 4000 is nice but it's not quite enough! I'll be really eager to see and keep track of all I learn through my environment with this approach.

I'm hoping my Anki review time on this deck will be a negligible amount per card either allowing me more time in my day or to be able to add a substantial amount more to acquire vocab faster. Of course reading lots of real Japanese will be review of it in context

An interesting idea. I'll give it a go after I finish KO2001.

Just added the 100 words i've collected over the last few days and reviewed them all. Took like 10 mins tops to review. That's what I'm after. Adding them took awhile because I was working inefficiently however a small adjustment to that and it's pretty quick to add cards too.

Speculation time...

30 words a day = roughly 900 per month. So in a whole year like 11,000 words.
50 words a day = roughly 1500 per month. Whole year = 18,000.
100 words a day = roughly 3000 per month. Whole year = 36,000 (dream on).

I plan on just putting in however many I want to/need to/feel is appropriate.

One thing I found interesting about KO was that, I personally wouldn't have gone out of my way to learn most of the words it taught me but I'm glad it taught me those words because no matter how big or small I do eventually wind up seeing them. Given I learned every new word in each sentence (95% of the time) It was a "leave no word behind" kind of approach.

The "leave no word behind" approach seems to have worked to retain about 80% of the knowledge. I wouldn't be opposed to taking it again given the results it got during KO but for now I'll stick to the stuff that is begging to be learned.

I’ve been using this method for awhile (nothing against those who like to sentences but it wasn’t working well for me---I was running into the same problems as mezbup). When I add new words that I come across in the wild, I put the example sentence in the answer portion of the SRS card since example sentences are good to have, just in my case not good as the question to SRS.

On the rare days that I have a ton of free time I can learn about 500 words (takes a few weeks to get the pronunciation down though so at first most of them get a rating of 2 or 3), normally I keep the daily added around 50-300 depending how many reviews I have. If I still can’t remember the pronunciation perfectly after a month, then I fail the card though for the majority of the vocab that doesn’t happen.

Also, I don’t learn them in order (except for Kanji Odyssey). For example let’s say you were learning 自体. You would put 自 in anki’s search bar and un-suspend any vocabulary that contains that kanji that you are comfortable with. Next you would do the same for 体, ect.

Overall stats:
correct mature cards= 97.9%
correct answers=88.1%

In the 1.8 months I’ve been using this method:

Finished Kanji Odyssey
Finished Core 2000
Almost done with Core 6000—only have 101 vocabulary words left
+ vocabulary learned in the wild
Grand total: 7635 :]

Since I focus on the pronunciation I’ve noticed my listening skills have steadily been increasing along with my reading ability [though that’s also partly because I turn on keyhole TV whenever possible.] Another cool effect is that I don’t need a J-->English dictionary to learn any new words.

Side note: completely agree with what others say about switching to a J-J dictionary as soon as you can, there are nuisances to some Japanese words that an English definition won’t catch.

Example of my average Anki card:
Front:
遂行
Back:
すいこう
物事をなしとげること。やりとおすこと。
「任務を―する」
キングダムハーツ2：任務を円滑に遂行するためにも事前の調査は欠かせないということだ

Hope you have as much luck with this method as I have Mezbup :]

I'd be interested to know whether you find that this approach is any worse than whole-sentences for helping you to remember usage details of words (eg 'is this verb transitive or intransitive?', 'which particles indicate what in the usual pattern for this verb?').

@fights: That's incredible! Thanks for you're very encouraging progress. It's the exact thing i'm looking for. I thought what I might do too is if I need it or it isn't too much hassle I could put an example sentence on the answer side so it's there if I want it. 7000+ words in 2 months is where it's at.

@pm215: I think it boils down to exposure to get usage details. One thing I found the sentence method didn't do for me that I thought it would have/should have was make me remember how to spit sentences back out. Just didn't eventuate. So in that respect no amount of usage details seemed to matter.

I don't even know what transitive and intransitive mean. I couldn't care less. I've started really seeing the big picture when it comes to learning Japanese now and have realised that you need massive amounts of exposure in different contexts and different types of material to truly come to grips with things. Aside from grammar you need to know collocations and Anki just doesn't have what it takes to teach you that.

http://www.robwaring.org/er/what_and_why..._vital.htm

So to get massive amounts of exposure you need massive amounts of vocab IMO. Almost like a catch 22. I've really taken what that essay has to say to heart. The figures kind of blow my mind at the number of words you need to read to encounter everything enough times to have really learned it all (so to speak). Anki can circumvent some of this but I don't believe it circumvents enough to actually warrant putting sentences into it anymore. The exception being sub2srs because these sentences are not "sentences" per se but rather "examples of human interaction in the target language" which can be fairly priceless

fightswumbrellas Wrote:Example of my average Anki card:
Front:
遂行
Back:
すいこう
物事をなしとげること。やりとおすこと。
「任務を―する」
キングダムハーツ2：任務を円滑に遂行するためにも事前の調査は欠かせないということだ

How do you deal with words that have multiple divergent meanings (eg 訴える)? Multiple meanings on the back and require that you get them all right to pass? Or do you add context to the front for that kind of word?

Basically I put as many definitions as my dictionary gives on the back. I don't have to get them all right to pass just have a firm sense of what the word means. Often you have a bunch of similiar meanings for any given word but sometimes a word has a couple of seperate meanings. When that's the case I generally want a firm sense of the different meanings and then i'll pass it. Besides, it's not a big deal if I do or don't know every meaning because if I see the word in the wild and what I know of it doesn't match up, I can search it in my deck and at that point i'll probably see one of it's alternate meanings and it will really click into place then.

Don't fret over every little detail. They will come in time.

edit:

my dictionary has a function where you can save words to a list and visit them later. There's 3 lists you can save to and each can have up to 100 words. I've decided what I'm going to do is use this functionality to separate new words I learn into the different lists depending on where I learn them from.

The three categories I'm going to track will be

A: Computer, Internet, Books, Magazines
B: Games
C: Music, Dorama, Anime, Movies

I'm interested to see after several months the data that this reveals to me in terms of where I'm learning the most from and what kind of words they are

I'm feeling you on the sentence thing. I don't know what I'm going to do, but something is going to give soon. I just don't have the patience to spend a bunch of time doing flashcards, and between KO2001, JRTK, and my normal Japanese dump/everything else deck, it's all killing me.

I'm interested in how you did KO2001 so fast. Did you use Anki along with it? Most Kanji comes with 3 example sentences, with (usually) different examples or reading/usage for each one. The way it is now, just adding 10 new kanji gives me about 30 new cards. Seems like it will take me forever to complete it this way.

I think you'll be fine. I'm a hardcore proponent for sentences, but as I've gotten more and more used to using Anki and seeing the benefits, I've realized that single words can work too. I think it's important to realize the pros and cons and be ready for them though. For a beginner, I think using anything but sentences is crazy since you learn SO much from them and in the beginning you will suck without a context so single words will actually take even more time. When you're already good though, you will often not need a context (contrary to popular belief). I have a few words in my deck I tried learning without context and while I wouldn't attempt to use them in my own sentences unless I was sure, I would easily understand them in context.

Tobberoth Wrote:I think you'll be fine. I'm a hardcore proponent for sentences, but as I've gotten more and more used to using Anki and seeing the benefits, I've realized that single words can work too. I think it's important to realize the pros and cons and be ready for them though. For a beginner, I think using anything but sentences is crazy since you learn SO much from them and in the beginning you will suck without a context so single words will actually take even more time. When you're already good though, you will often not need a context (contrary to popular belief). I have a few words in my deck I tried learning without context and while I wouldn't attempt to use them in my own sentences unless I was sure, I would easily understand them in context.

I think you're absolutely spot on here. Couldn't agree more.

Something fantastic I'm finding with this method already is i'm far more willing to take a new word that comes up and put it in my deck which is resulting in my acquiring more new vocab It's also pointing out to me how much I know already!

I'm kinda doing both, sentences and stand-alone words in Core2000. For every word that shows up in every sentence, I make a vocabulary card for it (if it doesn't already exist), and then I do the sentences themselves as well.

So far, it's going incredibly slow. Doing every word and then every sentence takes a bloody long time, and I really think I need to rethink my strategy. The thing is, I love the sentences for helping me with grammar, and getting used to different patterns, but man.. it just takes so long to review them. Every day I dread it - and I'm just talking about the reviews, not included adding new words.

brandon7s Wrote:I'm kinda doing both, sentences and stand-alone words in Core2000. For every word that shows up in every sentence, I make a vocabulary card for it (if it doesn't already exist), and then I do the sentences themselves as well.

So far, it's going incredibly slow. Doing every word and then every sentence takes a bloody long time, and I really think I need to rethink my strategy. The thing is, I love the sentences for helping me with grammar, and getting used to different patterns, but man.. it just takes so long to review them. Every day I dread it - and I'm just talking about the reviews, not included adding new words.

I'd stick with sentences until you have 2000 - 3000 words under your belt. From there it's easy enough to go without them because reading is do-able with a dictionary.

IceCream Wrote:hey well, you know i'm a proponent of words instead of boring sentences that have no use value... and it's great to be able to top up your vocab quickly, so go for it ：）

once you're through with all the core6k, here's where you can get soundfiles for JDIC entries thanks to Bombpersons:
http://forum.koohii.com/showthread.php?tid=4308

BUT: DON'T NEGLECT SENTENCES!

doing words alone is great for recognition, especially if you use audio. But, theres a lot more to language than vocab. Since the routes you've taken so far have all been through things like KO2001, there's an awful lot of sentence structures, interesting idioms, and applications of words that you really do need context to remember well. I definately advise mixing it up!

For example, i haven't got this set up yet, but i was planning to keep a word deck along with my sentences deck, because i don't feel like adding a sentence purely because it has a word in it i don't know. when i add a sentence, i'm more interested in something else about the sentence... a grammar pattern, or an unusual word application (for english), or a common phrase or idiomatic uses. When i went through one "my girl" episode, there were only 8 words i had not yet come across that weren't in a sentence i wanted. So, no big deal to just add those words. On the other hand, despite knowing all the words, i still took 40-50 sentences from the program. So... vocab isn't everything...

I totally recommend doing those production cards for sentences i talked about before, as well as listening cards (since you have to remember the sentence after it's finished to repeat it). That way, you will keep moving forward with natural uses, context, and vocabulary as well.

All very good suggestions I too feel the same way about sentences I've come to realise. Now that I have enough vocab to be able to understand plenty of native stuff i'm really starting to get common phrases and patterns and all the little things which make Japanese the colourful language it is.

If there's room for sentences in my deck it definitely belongs not to i+1 sentences but rather to "ah-ha!" sentences

Favourite phrase i've learnt recently いい加減にしろ！ｗｗｗ

Mezbup, it sounds like you're drowning in leeches. How often do you zap cards, either automatically with leech control or manually by deleting or suspending?

Using my leech settings, I expect to suspend at least 250 sentences by the time I reach the 3k sentence point. A 'no sentence left behind' approach would require more time during the learning stage and leave me with a fair number of lingering difficult items.

My experience is that words in isolation worked for me--for Esperanto. But, there's a critical difference between Esperanto and Japanese. I already knew most Esperanto grammar because it's similar to English or (failing that) Latin. Esperantists are, as a rule, very lax about usage and expect a diversity of styles. Japanese is of course different.

I may be wrong, and I hope so, but I think you're heading for a disaster. The "traditional" no-sentence approach consists of both words and grammar for a good reason: once you get beyond single-word exclamations in dictionary form, there are conventions for how words change and combine. I learn these conventions from three sources:

1) Definition.
2) Input.
3) Sentence review.

If you don't do sentences nor care to learn what transitive and intransitive mean, all of your grammar and usage eggs end up in the input basket. I see two problems here.

First, you can learn vocabulary from the exact same three sources:
1) Definitions.
2) Input.
3) Sentence review.

If you're not satisfied putting all your vocab eggs in the input basket, you must have some reasons. Don't those same reasons apply to grammar/usage?

Second, reviewing definitions instead of sentences leaves the root problem: leeches. If you're collecting hundreds of words a day, you're going to accumulate some hard-to-memorize ones that will sap your motivation and time unless you control them. Eventually, you end up in the same place now: mountains of reviews you feel you "have to do" that takes your time from the input you know you really need to do.

Solution: set your leech threshold low. If reviews take more time than you'd like, suspend failed cards rather than re-learning them.

Finally, re context dependent word recall:

Some words have multiple meanings, which you might just have to learn as multiple items. How many cards do you need to cover the range of meanings in「原稿を出す」「猫を外に出す」「車を修理に出す」「舌を出す」「旅に出る」「買い物に出掛ける」「庭に出る」「出る釘」etc, etc?

The dictionary says 「出る：……」and「出す：……」 and maybe 「出掛ける：……」. But, who cares about the dictionary? When you can't remember a word in a new context, that just means that you have a new definition to add to your mental dictionary. Whether or not the official dictionary says it's a new word is entirely beside the point.

Make your internal dictionary fit your own mind and needs--not what some lexicographer put on paper. Stop reviewing hard items. I believe you must do those two things whether you study sentences or isolated words. If not, either approach will be ineffective.

I, too, started doing sentences from Core2000 at first, but let's face it: the sentences in the Core-series are pretty boring. And while you'll have a sentence to provide some context for new vocabulary, you are still missing a context for the sentence itself.

Next, I tried just SRSing vocabulary from the Core2000 and Core6000, but got tired of that too, mostly because the same lack of context and the fact that I felt I was trying to learn a lot of vocabulary that wasn't so useful to me (at least not yet).

What I do now is reading real material in Japanese (in my case: game reviews, manga, essays, short stories) and add each and every new word that I encounter into my deck, by hand. Having a real context in addition to the dictionary entry usually gives me a very precise idea about the meaning and nuance of the word (compared to the Core-sentences). I've also noticed that I remember the context where the word was from even after several months, despite the fact that I don't usually include the source sentence in the SRS.

Also, I generate both recognition and production cards from every word. I have a separate model for words that contain kanji and pure kana-words.

Kanji words go from kanji -> reading + english and reading + english -> kanji. For kunyomi words, I have an additional field in the question when doing production that lists all the variants that I'm NOT supposed to draw this time.

Kana words go from word + example sentence -> english meaning and vice versa.

Since I practice kanji drawing with the production cards, I've stopped doing all Heisig reviews altogether. At this point, I don't see any merit in further enforcing the Heisig keyword to kanji links. I might forget some of the more rare kanji before I encounter them in the wild, but I'm confident that they will be easy to re-learn after getting them as part of real vocabulary for the first time.

I've been doing sentences from the smartfm. KO list, and turning them into additional production cards after they reach a due date higher then 2.5 months. After about 500 sentences with the "learn every word in the entire sentence approach" (not sure how many words I learned as some sentences had 4+ new words) I added the single word version of the list. I just started it, so it's mostly words I learned that from sentences already (though some are new as I usually don't add the ones without audio - some of them are ridiculously LONG) - The idea I had was to kind of fill in the cracks, and better learn the root word, as I always focus on/hear conjugations in sentences. The single word lists go super fast and I can do the production version pretty easily too.

I really like some of the idea's here - too many good ones to point to just one poster, but I found myself nodding my head as I read through - I'm experiencing the same problems with sentences - I also see the benefits too, but I think the more dynamic the strategy the better. Though I know I'm too much of a beginner to have good conversation, I have found that I sometimes have to think of a sentence from my deck and mentally find a word that is in the middle of the sentence from the sentence pattern to say it - that's just too slow and annoying - and was one of the reasons I decided to also study words in isolation. I think I will alter my approach as I go.

wildweathel Wrote:Mezbup, it sounds like you're drowning in leeches. How often do you zap cards, either automatically with leech control or manually by deleting or suspending?

Why would you ever suspend a card? You put that card in your deck for a reason, and even if you don't like it, that word or sentence isn't going to magically disappear from the language. I think if I kept my leech settings on, 90% of my deck would end up suspened, because I randomly forget crap all the time.

strugglebunny Wrote:You put that card in your deck for a reason,

Yes. And that reason is: "it would be nice not having to look this up," not "I must learn this now or fail Japanese forever."

Quote:that word or sentence isn't going to magically disappear

It's a question of priority. Reviewing hard items while easy ones wait isn't going to grow my vocabulary. If that word or sentence is common, seeing it repeatedly will make it easier and I'll be able to pick it up later. (Thus, why I suspend rather than deleting.)

Quote:90% of my deck would end up suspened

Actually, with fairly aggressive settings (6 fails = suspend), it's less than 10%, closer to 5%. You may have to tune, but I do recommend trying it.

It's pretty natural to rely less and less on sentences as you progress and build your mental corpus, because you can just slot this stuff into what you already know (for words being able to mentally contextualize it pretty easily based on the original encounter). Personally though, after 'perfecting' (ongoing) my own approach, I find it harder to do words out of context. For me, more elaborate/complex = easier and more versatile. Tangent: Just noticed another good reason to convert 'usually kana' words to kanji, if only for the first pass-throughs and even (esp.) if you already know those readings: I use the kanji's extra mental hooks to help remember the semantics/sounds, so more complex = synergistically easier. (The main 'reading' model for Japanese in cognitive science is something with multiple routes operating in parallel, where you've got the kanji using heavy visual-semantic access independent of yet in conjunction with phonetic-semantic and visual-phonetic. BTW I've been reading how the brain uses 'imageability' [easily visualized definitions, usually concrete] for low frequency two-kanji kun'yomi readings by depending more on the visual-->semantic-->phonetic route, interesting stuff, should take advantage of those basic brain dynamics... That's actually how I learn most readings hehe, with the more unfamiliar stuff requiring a heavier touch.)

At the moment, I think of it in terms of a general, mostly fixed 'reference corpus' (Costas Gabrielatos on 'types of corpora'), and a specialized, dynamic 'monitor corpus'. Specifically, stuff like smart.fm/grammar cards/KO2001 for the general stuff, and subs2srs for video clip watching/an extemporaneous deck I add to, well, extemporaneously as I encounter interesting things that I want to elaborate on if I think it won't stick as well without SRSing.

Well, for new words I encounter in the subs2srs video clip decks, I don't learn the words or grammar points there, I make a list of the new ones and reference them in the reference corpus, or add them as single words/cards that have multiple complementary sensory dimensions ;p. The latter being kind of a like a subdeck in the monitor corpus... Thus I can focus on usage/parsing/listening/etc. for the original 'monitor' decks. Eventually it all begins to collapse* into 'just SRS a small percentage of stuff I come across outside the SRS' pretty organically, in my mind, but I still think those basic memory structures are best built and maintained through sentences and then through the smallest multidimensional 'units'/cards you find most effective.

Eh, sorry for the ramble, your enthusiasm infected me. ^_^

Oh! And same for 'sensorimotor' memory, been reading how those aspects of memory are triggered even by just mentally rehearsing it, how spaced repetition applies to that kind of memory as well, and how important a role it plays esp. in kanji alongside visual memory. A while ago some people were concerned about the role of 空書 (writing kanji in space) in the computer age but I think we've learned to incorporate that in RTK and so on in a more minimal yet still effective way with SRS.

Really going to have to sort my references and links for this stuff. Fascinating reading.

*Well, before that point, I'm planning on a strategy that combines the best of both worlds for a while, but that's another post...

First of all, I'd like to say that suspending difficult cards in order to fit in more reviews of easy cards is perhaps the worst idea I've ever heard.

mezbup Wrote:...the main barrier to understanding Japanese is vocab.

Every beginner should be required to get this tattooed on their forearm.

A lot of people seem to have forgotten that Khatzumoto (the reason everyone is into the sentence thing... remember him?) started out with extremely basic 2-3 word sentences such as 結婚の相手 and 化け物が出た. These are effectively vocab cards.

Context is important, yes, but nowhere near as important as actually knowing the word in the first place. Vocab decks allow to you to get through thousands of reviews each week. You can only review thousands of sentences each week if you're doing i+1, which is impossible for a beginner. Grammar is so much easier to learn when all the words in a sentence are familiar.

And another added benefit of learning vocab quickly and in isolation is that native material becomes much more accessible much more quickly.

fightswumbrellas Wrote:In the 1.8 months I’ve been using this method:

Finished Kanji Odyssey
Finished Core 2000
Almost done with Core 6000—only have 101 vocabulary words left
+ vocabulary learned in the wild
Grand total: 7635 :]

It's a shame JLPT1 isn't in 1.2 months. You could be The One.

Ah, an overload of arbitrary and overly narrow-minded statements about language-learning blended with some juvenile tones. My forum-OCD is acting up again, makes me want to smash inferior arguments. Invocation to Khatzumoto was the trigger, I think.

But I shall refrain. Posting too much. ;p

Having said that... No disrespect. ^_^

Yes, advocating a balanced approach and offering an alternative to blindly following Khatzumoto, how narrow-minded & juvenile...

I would like to join the oncoming debate between harhol and nest0r, but unfortunately I can't because I seriously have no idea what harhol tried to say in his post.

That we should use sentences? That we shouldn't? He clearly said suspending leeches is a bad idea, but he never in any way said why, so it's hard to argue against that. He also said that learning vocab in isolation has the benefit of letting you use native material early. Unfortunately, there was no reasons added here as to why it's a benefit compared to sentences.

So all in all, I have nothing to add because.... I don't know where to add it.

Tobberoth Wrote:I seriously have no idea what harhol tried to say in his post.

tl;dr alert



The standard RevTK advice for RTK1 & Tae Kim graduates is to learn several thousand complete sentences from smart.fm or KO2001 in order to get a ready-made, audio-assisted head start on Khatzumoto's 10,000 sentence method. People are told to avoid learning grammar or vocab in isolation because, without adequate context, the words and rules they learn will apparently be of little practical use.



Here are some things I have observed through personal experience with this method:

1) The grammar used in smart.fm's sentences (can't speak for KO2001) is extremely basic and of questionable use to a Tae Kim graduate. You learn context, yes, but stiff & unnatural context. And I can't recall coming across any grammar patterns in smart.fm which I hadn't seen covered in Tae Kim's guide. Since the sentences are so basic and cover grammar which you already know, it seems silly to spend hours each day reviewing them.

2) It bypasses the i+1 methodology, which is the entire foundation of the 10,000 sentence method and one of the main reasons why it has the potential to be so effective. Most of the first few hundred core 2000 sentences are i+2 or i+3, with some being as high is i+6. These sentences slow you down, frustrate you when you fail them and sap your motivation. Then, as you get further into smart.fm, you start to come across dozens of useless i+0 sentences.

3) There's a good chance that you will learn certain words only in the context of a sentence and will therefore be unable to recognise them elsewhere. Maintaining an isolated vocab deck not only eliminates this problem, it also allows you to be sure that you know the standard/dictionary form of the word and not just a particular tense/usage.

4) In the time it takes an average person to finish studying all core 2000 sentences, they could instead have learned the core 6000 vocab in isolation. They could then start to explore native media with a vocabulary roughly three times a big.

5) Above all, it is incredibly slow & tedious. I would find myself wanting to do anything but return to studying & reviewing smart.fm sentences.



With that in mind, here are the reasons why I believe studying vocab in isolation before commencing AJATT (the IceCream method?) would be a more useful approach:

1) Nouns. There are thousands upon thousands of daily-use nouns which you simply need to know and which don't change with context.

2) Reviewing vocab is quick & easy. You can burn through hundreds of reviews in an hour and thousands of reviews each week. Also, failing a word isn't nearly as demotivating as failing a sentence. Starting with smart.fm or KO2001 is like jumping into the middle of the ocean and slowly swimming toward the shore, whereas starting with vocab in isolation is like starting at the shore and swimming out into the ocean.

3) The more vocab you know, the easier it is to study/memorize new grammar concepts. If you're attempting to understand the different subtleties of a particular expression, the last thing you want to be doing is wondering what the words either side of it mean.

4) The more vocab you know, the easier it is to study/understand/enjoy native material and the sooner you can begin that phase of your studies.



With that in mind, here is what I would consider to be the ideal approach for a complete beginner who has just finished RTK1 and Tae Kim:

1) Core 6000 vocab in isolation. 2-3 months.

2) Sentence mine All About Particles (in true i+1 fashion since you already know all of the vocab). ~300 sentences; 3-4 weeks.

3) Sentence mine A Dictionary of Basic Japanese Grammar (in true i+1 fashion since you already know all of the vocab). Don't add every single sentence, just enough so that you understand each entry and its variations clearly. ~300 sentences; 3-4 weeks.

4) Commence AJATT in true i+1 fashion while maintaining and adding to your vocab deck.



I hope you can see from this what my overall POV is and why I'm in favour of studying vocab in isolation. Khatzumoto's methodology is sound and he is undoubtedly an inspirational character, but I believe that ready-made solutions like smart.fm and KO2001 have caused some people to lose sight of the foundation of his method, which is i+1 learning. When you're a beginner, the only way to do i+1 is to study vocab in isolation. And, as I said in my previous post, it's clear from the starter packs that Khatz has posted that he was basically studying vocab to begin with anyway. As you become more advanced, studying vocab in isolation remains useful, because while there are many words which are hard to pin down without adequate context, there are just as many which mean one thing and one thing only. Besides, if the definition is ambiguous, there's nothing to stop you from having example sentences in the answer section detailing the various uses.



Oh and here are the reasons why I believe that suspending difficult cards is a bad idea:

1) You might not need to know it right now, but you'll need to know it eventually.

2) If you keep suspending difficult cards, you'll eventually end up with a mountain of difficult cards which you'll need to tackle all at once. Then you'll be even less likely to want to study them!

3) A particular difficult card might be way more important/useful than a particular easy card. Yes, you may have successfully reviewed 頭寒足熱 50 times, but that's not very helpful if you've suspended words like 働く and 意見.

4) It's a cop out. "If it's difficult, I'll just skip it" is not a good mindset to have.



*faints*