+- kanji koohii FORUM (http://forum.koohii.com)
+-- Forum: Learning Japanese (http://forum.koohii.com/forum-4.html)
+--- Forum: General discussion (http://forum.koohii.com/forum-8.html)
+--- Thread: The key to fluency: 10,000 words in one year. (/thread-10466.html)
The key to fluency: 10,000 words in one year. - uisukii - 2013-03-05
Rael89 Wrote:Wow. amtrack's thinking is almost as dated as the crummy trains company he runs.Yeah... be that as it may or may not be, but flame baiting is just as or if not, worse than an honest opinion which may or may not be overly sound or evidential supported.
The key to fluency: 10,000 words in one year. - Stansfield123 - 2013-03-05
nadiatims Wrote:It's not about the complexity or interconnectedness of a given word within the language. It's a matter of the interconnectedness of the word with other memories in your brain making it instantly and effortlessly retrievable.My point is that word and concept are not the same thing. The primary purpose of one's native language is to provide you with the symbols that allow you to classify and organize concepts (give you the symbols which allow your conscious mind to "handle" those concepts).
However, the primary purpose of any language after that first one is to allow you to communicate with people using a different set of symbols. When you learn a second language, you don't need to integrate and organize a whole new set of concepts, you already have that in place. All you need to do is connect the new symbols to the existing concepts.
That does not require the kind of mental processes (of integration - or induction, if you wanna get "technical" - process which, on a physical level, results in neural interconnectedness) learning one's first language requires.
corry Wrote:What I got from reading Kraschen is that if you are trying to pick up a language on your own then the way to do it is to make sense out of the Japanese you are reading and listening to. I dont think it really matters how you make sense out of it. But SRS'ing random bits and pieces as a way to prep yourself beforehand doesnt seem like the most direct way to me.No, SRS-ing random bits and pieces wouldn't. SRS-ing basic grammar points and basic vocab presented in an i+1 formatted sentence, to add context to it, however, might.
Yes, theoretically, you could use what Stephen Kraschen calls "comprehensible and interesting" reading materials to learn Japanese right after you learned the Hiragana. In fact, that would be awesome. However, in practice, that's impossible, because interesting Japanese writers aren't going to write for an audience of a few hundred language enthusiasts familiar with Krashen's theories. The next best option is using SRS, subtitled television and movies, grammar points, textbooks, etc. to get to a level where interesting reading material becomes comprehensible.
The key to fluency: 10,000 words in one year. - Stansfield123 - 2013-03-05
amtrack Wrote:There are very very intelligent people in charge of learning, and teaching at universities.No, there aren't. The state of modern education proves that there aren't.
To your overall point, SRS software clearly does not artificially space out cards in a supposedly correct fashion to optimize learning them. If that's what it did, you wouldn't have those fancy buttons at the bottom, which allow the user to CHOOSE the next time he intends to review the card. You'd have one button, which reads "Next", instead. So your whole argument is a giant straw man.
The purpose of SRS is to minimize wasted time by allowing users to easily schedule the next time they study a card based on how well they know it. SRS algorithms allow a user to choose between an interval anywhere from one minute all the way to months, in the vast majority of cases (the few exceptions are the new cards, when the options are limited to between a minute and several days, instead). This couldn't be further from the straw man you constructed and are arguing against, of a pre-determined schedule of reviews.
What you seem to be arguing for, instead, is completing RtK by going over the character for One the same amount of times as going over the character for Attack or Entrails. You really think that's just as efficient as putting more time into difficult Kanji, and less into easy ones?
amtrack Wrote:But once you "learn it," forgetting it is extremely difficult. All I'm saying is that the SRS algorithm itself does absolutely nothing for you.Damn it. This is what happens when you scan instead of read, kids. I completely missed this bit, and its implications, before I typed out the long response above.
The notion that forgetting isn't something a teacher/learner should account for is not really worth dignifying with a response.
The key to fluency: 10,000 words in one year. - Irixmark - 2013-03-05
Stansfield123 Wrote:Second that. I spent a few hours in the educational studies library and was appalled by the quality of the "studies" of learning. Evidence-based approaches to teaching are rare. But things seem to be changing with more randomized controlled trials in education as well. Just that they never, ever seem to feed back into teaching...amtrack Wrote:There are very very intelligent people in charge of learning, and teaching at universities.No, there aren't. The state of modern education proves that there aren't.
For what it's worth, I tried using iFlash instead of Anki, which doesn't have an algorithm but just allows you to set intervals depending on how often a card is rated as correct. I didn't get nearly the retention rate I got with Anki, so for me at least the algorithm seems to do something.
The key to fluency: 10,000 words in one year. - nadiatims - 2013-03-05
Stansfield123 Wrote:However, the primary purpose of any language after that first one is to allow you to communicate with people using a different set of symbols. When you learn a second language, you don't need to integrate and organize a whole new set of concepts, you already have that in place. All you need to do is connect the new symbols to the existing concepts.I get that. But I'm not talking purely about understanding the concepts (word meanings), I'm talking about having them become automatic which will only start happening when you start actually using them as a unit of meaning to understand unknown messages in context not just a=b flashcards. You could theoretically cram 20k vocabulary in anki or something but you'd still have trouble reading a book because none of that will have become automatic, nor will you have any grammar intuition. The point about SRS is to realise that unless you set it up to emulate reading a book (why bother?) all it installs in your brain is a link to a word in your L1 (or an image or definition or whatever you use). It's basically just like tying your shoelaces before a race. It can be useful but how many times do you really need to do it? Just do it once and then start running. If they come undone tie them up again, or just run barefoot. Maybe not a great analogy but I think you get my point. The real learning happens after that initial dictionary lookup or flashcard review when the word is encountered again and again (ie. practiced in reality) and used to interpret or convey unknown messages.
That does not require the kind of mental processes (of integration - or induction, if you wanna get "technical" - process which, on a physical level, results in neural interconnectedness) learning one's first language requires.
Stansfield123 Wrote:To your overall point, SRS software clearly does not artificially space out cards in a supposedly correct fashion to optimize learning them. If that's what it did, you wouldn't have those fancy buttons at the bottom, which allow the user to CHOOSE the next time he intends to review the card.This is the big problem with SRS and where it really wastes a lot of time. It helps you master a task that has little relation to actual language use and that is of rather limited usefulness. Namely it teaches you that a=b and it tests you on your ability to remember that same fact a few minutes later, 30 minutes later, a day later, a week later etc. As if this mattered. Who cares if you remember something a minute, day or even a week later? None of that matters. All that matters is that you are able to use it in the long term. SRS is deceptive because it sets up and helps you accomplish a task (jumping through arbitrary "spacing" hoops) that is ultimately unnecessary.
The key to fluency: 10,000 words in one year. - headphone_child - 2013-03-05
nadiatims Wrote:This is the big problem with SRS and where it really wastes a lot of time. It helps you master a task that has little relation to actual language use and that is of rather limited usefulness. Namely it teaches you that a=b and it tests you on your ability to remember that same fact a few minutes later, 30 minutes later, a day later, a week later etc. As if this mattered. Who cares if you remember something a minute, day or even a week later? None of that matters. All that matters is that you are able to use it in the long term. SRS is deceptive because it sets up and helps you accomplish a task (jumping through arbitrary "spacing" hoops) that is ultimately unnecessary.What's wrong with saying that both approaches (SRS and consuming native material) can be useful? I dove into native material immediately after RTK and picked up words that way for a while, but when I started getting diminishing returns I started doing Core decks as well. It's true that SRS only gives you the a=b which is not what you want in order to become fluent, but I think it's a good start. It was those words that were in Core that, when encountered in native material, stuck with me both quickly and lasted. If I didn't do Core, my progress probably would have been slower as it would have taken more encounters of the same word before I'd start remembering it.
I do think that SRS and nothing but SRS is bad, but I don't think anyone was arguing otherwise. It's just a matter of how much SRS vs native material, and when. SRS felt more effective in the earlier stages of learning for me (though I don't do much SRS anymore, so it may be my confirmation bias that makes me hope it doesn't matter as much now), but what's best would depend on the person for sure.
The key to fluency: 10,000 words in one year. - uisukii - 2013-03-05
nadiatims Wrote:That is only really applicable if the individual doesn't use an SRS in conjunction with experiencing the target language through multiple epistemological points of connection. Find a person who isn't also reading more native material as their vocabulary and pattern comprehension expands, or isn't using their SRS experiences in conjunction with a range of other media in their target language.Stansfield123 Wrote:To your overall point, SRS software clearly does not artificially space out cards in a supposedly correct fashion to optimize learning them. If that's what it did, you wouldn't have those fancy buttons at the bottom, which allow the user to CHOOSE the next time he intends to review the card.This is the big problem with SRS and where it really wastes a lot of time. It helps you master a task that has little relation to actual language use and that is of rather limited usefulness. Namely it teaches you that a=b and it tests you on your ability to remember that same fact a few minutes later, 30 minutes later, a day later, a week later etc. As if this mattered. Who cares if you remember something a minute, day or even a week later? None of that matters. All that matters is that you are able to use it in the long term. SRS is deceptive because it sets up and helps you accomplish a task (jumping through arbitrary "spacing" hoops) that is ultimately unnecessary.
Essentially, it seems as though you are arguing your points against a constructed unrealistic target, which only becomes applicable if people were using SRS software as a part of their target language study in a bubble; 'srsing' for the sake of getting better at 'srsing'. This is something brought up a few pages back now, which has seemingly been ignored.
One of the large problems, in my opinion, with these discussions, is that someone forms an argument reliant on a hyperbolic position, and others take the bait and defend something which they aren't really proponents on initially. Experiencing target language through native media, as opposed to SRS, as though it were a binary position, does not exist.
The key to fluency: 10,000 words in one year. - Stansfield123 - 2013-03-05
uisukii Wrote:That is only really applicable if the individual doesn't use an SRS in conjunction with experiencing the target language through multiple epistemological points of connection.Ok, so then what role does the SRS software perform? If by learning how to say a concept in Japanese, you haven't integrated that Japanese word into your context of knowledge, then what does SRS do exactly? Wouldn't it be just as easy to just form those "multiple points of connection" just by experiencing the target language?
Nadiatims' contention is that SRS simply associates two random symbols. Mine is that it associates a random symbol with a concept, thus making that symbol mean that concept (along with every concrete that concept stands for). My contention is that if I were to use Anki to really drill 100 simple Japanese words into my head, and then started speaking English, except that I would use those 100 Nihongo kotoba naka place of Eigou kotoba itsudemo applicable, watashi dekiru suru kore, iie problem, soretomo anata dekiru read kore, iie problem.
There would be no further "connections" that would need to be filled in. I'm sure it's harder for you to read the above sentence, but not because you're searching for some missing connection to the concept, it's because you're having trouble recognizing the visual shape of the poorly written romaji (but read it two or free times, and note how that problem goes away - even though you clearly haven't seen those words in any more contexts, since ten seconds ago - and the sentence becomes no less accessible than any other English sentence you may read - it is in fact an English sentence, it's just encoded using a very primitive method).
So my contention is that learning just the words associated with concepts (whatever means you're using, I certainly wouldn't use Anki to learn most of my words, but I could if I wanted to) means you're all done with those words (well, those words in one specific sense, not with the many different, nuanced meanings of the same word). My contention is that, when you finally start using those words, the blanks you're filling in have nothing to do with further understanding/integrating the word. They have to do with grammar, forming visual memory when reading the words, learning the many nuances (which, unlike for learning just one meaning, you DO need many contexts for), forming auditory memory (which is more than just the visual memory of a single word, auditory memory involves remembering how entire expressions sound) to gain listening comprehension, and finally pronunciation.
The key to fluency: 10,000 words in one year. - darkauras - 2013-06-21
I'm gonna interrupt the age old SRS debate here and just give my thanks for this thread. Although everything said in the OP's post was completely obvious after reading it (I mean...when you stop to think about it it makes complete sense) it really gave me some perspective. I've started doing a dedicated 35 words a day (I do fail to meet this mark every once in a while) and I've improved by leaps and bounds. My biggest weak point has always been vocab rather than grammar, and this is helping me overcome that.
I just had to come out of hiding to say thanks. I shall now return to lurking and the debate can resume...
The key to fluency: 10,000 words in one year. - RawToast - 2013-06-21
Quote:I have set it to 200 without any noticeable consequences, except that my reviews quickly went down to ~100 per day to ~70, but now they're back to ~90 again, because I encreased the number of cards added per day.Stian, Matsu, and others who have increased the default spacing, are you only doing one or more cards per fact? I am just wondering how this works, as I am looking to push through the core 6k at a much faster pace after July/JLPT.
To me, the spacing is helpful because it limits the number of flashcards I have to review every day. I only review about 100 of my 3000 or so cards each day, and my 90% retention seems to "prove" that it works for me.
It does require a lot of immersion to work proberly though, and I never "learn" anything during my anki reviews.
The key to fluency: 10,000 words in one year. - Stian - 2013-06-21
The major difference is that the easy cards get out of the way much faster whereas the hard ones still remain hard. Of course, the retention rate will drop (between 80-90 for me instead of 90+)
I'm not doing core, however, so my experiences might differ. The major advantage however, is the reduced number of reviews. I'm not sure how much time you save though, some people (dtcamaro, I think?) claim that you should have 20x as many reviews as the daily amount of new cards, but I never experienced that (it was more like 8x-10x) even before I altered the intervals. Now, they're about 5x-6x...
Sorry for the messy post.
Just go ahead and try it for a few months; if you mess up, you can always reset the settings.
To the OT: The OP completely ignores that he focused on grammar for an entire year before cramming loads of vocabulary. Getting a good grasp of grammar and learning 10k words in two years sounds more reasonable to me...
The key to fluency: 10,000 words in one year. - nadiatims - 2013-06-21
Stian Wrote:To the OT: The OP completely ignores that he focused on grammar for an entire year before cramming loads of vocabulary. Getting a good grasp of grammar and learning 10k words in two years sounds more reasonable to me...How thorough does that understanding need to be though? and how thorough can it be if you don't yet have the vocabulary to do serious reading and listening.
The very basic grammar that is broadly applicable and useful to beginners probably needn't take more than a couple of months to get to grips with. Stuff like:
- how to get to dictionary form (verbs and adjectives)
- recognizing the basics of how ha/ga/wo/ni give a words function in a sentence.
- knowing the words rareru, saseru and how they relate to the above.
- recognizing subordinate clauses
- understand that how te leaves a clause un-finalized
- to/mo
- ra/ba
Pretty much everything else (and much of the above in fact) can be figured out using dictionaries/glossaries and/or content with translations. I'm pretty sure I learned verb conjugation as a natural by-product of deconjugating words in order to check them in dictionaries.
There are so many grammar patterns, exceptions and things that just are that beyond those basics I think you have to pick it up little by little.
It's often said that really learning of Japanese starts once you're around N1 level. I think a lot of learning of grammar and finer nuance only really starts when the overall meaning of most japanese has become transparent because you've acquired most of the vocabulary. Having the vocabulary down really frees up a lot of brain processing power to start noticing a lot of the nuances of how words are used.
The key to fluency: 10,000 words in one year. - Animosophy - 2013-06-21
nadiatims Wrote:It's often said that really learning of Japanese starts once you're around N1 level. I think a lot of learning of grammar and finer nuance only really starts when the overall meaning of most japanese has become transparent because you've acquired most of the vocabulary. Having the vocabulary down really frees up a lot of brain processing power to start noticing a lot of the nuances of how words are used.I agree. I'm at an awkward phase in my learning where basic grammar and sentence structures are "farmiliar" to me, but I severely lack reading and listening comprehension because both my knowledge of grammar and vocabulary aren't so firmly rooted in my mind yet. I need more time to do what I'm already doing.
I suspect that near-native reading and listening comprehension speed is necessary to able to think in Japanese with some level of consistency. It's a bit crap having to wait for that point, when everything clicks. Hoping 12 months more is enough time, and mezbup's article is encouraging
The key to fluency: 10,000 words in one year. - s0apgun - 2013-06-21
darkauras Wrote:I'm gonna interrupt the age old SRS debate here and just give my thanks for this thread. Although everything said in the OP's post was completely obvious after reading it (I mean...when you stop to think about it it makes complete sense) it really gave me some perspective. I've started doing a dedicated 35 words a day (I do fail to meet this mark every once in a while) and I've improved by leaps and bounds. My biggest weak point has always been vocab rather than grammar, and this is helping me overcome that.Nice update. Glad to hear this method is working for you. You've provided me with additional motivation.
I just had to come out of hiding to say thanks. I shall now return to lurking and the debate can resume...
The key to fluency: 10,000 words in one year. - Aspiring - 2013-06-21
as an aside, k Wrote:Is the SRS alone enough? I want it to be. Fundamentally, I believe that every large problem can be solved through good systems…A good system gives us a way to connect tiny local actions into a larger global goal or solution. But in my experience with and observation of purely SRS-centric, low-immersion language learners, I have yet to see good results. I have seen people spin their wheels just dry-SRSing themselves into oblivion, avoiding immersion, with its rough edges and frequent lack of certainty, like a drunk salaryman on the train. I hesitate to hypothesize, but I think it’s safe to say that high-concentration, high-quantity exposure to engaging native materials is a far better overall predictor of fluency than SRSing.
The key to fluency: 10,000 words in one year. - dizmox - 2013-06-21
Animosophy Wrote:It's never too early to think in broken Japanese, though.nadiatims Wrote:It's often said that really learning of Japanese starts once you're around N1 level. I think a lot of learning of grammar and finer nuance only really starts when the overall meaning of most japanese has become transparent because you've acquired most of the vocabulary. Having the vocabulary down really frees up a lot of brain processing power to start noticing a lot of the nuances of how words are used.I agree. I'm at an awkward phase in my learning where basic grammar and sentence structures are "farmiliar" to me, but I severely lack reading and listening comprehension because both my knowledge of grammar and vocabulary aren't so firmly rooted in my mind yet. I need more time to do what I'm already doing.
I suspect that near-native reading and listening comprehension speed is necessary to able to think in Japanese with some level of consistency. It's a bit crap having to wait for that point, when everything clicks. Hoping 12 months more is enough time, and mezbup's article is encouraging
I don't ever remember thinking in English, not even as a beginner.
The key to fluency: 10,000 words in one year. - Eikyu - 2013-06-22
darkauras Wrote:I'm gonna interrupt the age old SRS debate here and just give my thanks for this thread. Although everything said in the OP's post was completely obvious after reading it (I mean...when you stop to think about it it makes complete sense) it really gave me some perspective. I've started doing a dedicated 35 words a day (I do fail to meet this mark every once in a while) and I've improved by leaps and bounds. My biggest weak point has always been vocab rather than grammar, and this is helping me overcome that.Well, you're right. Mezbup (OP) was right. Japanese is a numbers game. Once you learn enough vocab, the rest is not too bad. If you get a jumpstart through SRS, that will help you tremendously. Don't get depressed by high review counts though. With 35 new cards per day, you will have a lot of reviews to do and might have to review less some days.
I just had to come out of hiding to say thanks. I shall now return to lurking and the debate can resume...
The key to fluency: 10,000 words in one year. - Sheriff - 2013-06-22
nadiatims Wrote:This is the big problem with SRS and where it really wastes a lot of time. It helps you master a task that has little relation to actual language use and that is of rather limited usefulness. Namely it teaches you that a=b and it tests you on your ability to remember that same fact a few minutes later, 30 minutes later, a day later, a week later etc. As if this mattered.But isn't a=b perfect for learning readings? I think it is. And I think SRS is a powerful tool to tackle this big obstacle when it comes to learning Japanese or Chinese. And while your at it, get one or two meanings in there. What's the harm? Just don't fool yourself into thinking SRS will teach you the full meaning of an L2 word with all its nuances and proper usage depending on the context. I guess this is one of your points and I fully agree. There's no way around consuming lots of native material to get that done.
nadiatims Wrote:Who cares if you remember something a minute, day or even a week later? None of that matters. All that matters is that you are able to use it in the long term.How can you use something long term when you can't even remember it after a week or a month? This bit is highly contradicting, no? Or I probably just missed the point. ^^
Y'all have a nice weekend!
The key to fluency: 10,000 words in one year. - Animosophy - 2013-06-22
dizmox Wrote:It's never too early to think in broken Japanese, though.Any tips on how to make the most of broken Japanese?
I don't ever remember thinking in English, not even as a beginner.
I had the idea of keeping a child-like free writing diary. I'd write like a 9 year old Japanese kid, basically, and I suspect my entries will get progressively more complex. Hmm... now that I've said it, I have no idea why I didn't just start doing it.
Still, I'd appreciate any experiential advice you could offer.
The key to fluency: 10,000 words in one year. - s0apgun - 2013-06-22
^ http://lang-8.com/ is a great resource for beginners and experts to practice thinking in japanese. the trick is keeping up with it like anything else
The key to fluency: 10,000 words in one year. - dizmox - 2013-06-22
Animosophy Wrote:I guess I played around with Japanese in my mind, saying random stuff to myself. No idea if that's of any use though.dizmox Wrote:It's never too early to think in broken Japanese, though.Any tips on how to make the most of broken Japanese?
I don't ever remember thinking in English, not even as a beginner.
Are other people subvocalising stuff they want to say in English then try to translate it? I don't really subvocalise any of my thoughts (unless I'm imagining actual conversation), so I have no idea, but that doesn't sound like a good habit.
The key to fluency: 10,000 words in one year. - NightSky - 2013-06-23
I don't want to spend too much time reading the arguments here, just wanted to add my thoughts that Mezbup mostly has this right and its been quite eye opening to me. When I first got to Japan a foreigner told me "Learn 4 words a day and you will be fine".
Maybe its just the process of sitting down to learn 4 that might make you study more, but realistically only 4 a day will take forever to get to a good level. When I try to really focus on Japanese I try to add up to 100 new words or phrases a day into Anki from things I've read. I have to read for probably 4-5 hours to even find 100 words I don't know, but even in a couple of weeks of this I notice substantial improvements.
So far this year with everything else going on in my life I've added about 5000 new words to my Anki deck. I'm hopeful I can keep going but its getting harder and harder to even find words I don't know to put into Anki, so its becoming the case now that Anki is becoming less and less useful and its all about native materials and exposure.
For Chinese I don't find it possible to add this many because the words dont stick. With Japanese I knew a lot already and spoke fine so adding 100 more a day is okay, because with some of them I only ever see them once and can answer 'easy' every single time after. With Chinese it doesnt stick and I quickly get snowballed under reviews trying to continue at that sort of pace.
The key to fluency: 10,000 words in one year. - Animosophy - 2013-06-23
^ you could download the word frequency list here http://forum.koohii.com/showthread.php?pid=69977#pid69977 and use something like http://www.tracemyip.org/tools/remove-duplicate-words-in-text/ to remove from the list all the vocabulary you've got stored in Anki. This is what I figured I'd end up doing in your position lol.
The key to fluency: 10,000 words in one year. - Livonor - 2014-03-05
Ah the classical non-SRSxSRS discussion, I like how it gives a familiar and traditional feeling to the thread. After leaving this habits for the sake of my sanity I promise never engage in a long discussion again, and would be good to keep that promise, but I always feel that facts spoken launder than arguments, because for every argument there's a counter argument so nobody is never wrong.
I've been learning 400-1000 words a day, if that's not being efficient I don't know what is. Of course, I don't really know them, at least not by many other's people standards, but after a couple of months I will.
Anyway, sorry for resurrecting the thread
The key to fluency: 10,000 words in one year. - Guoguodi - 2014-03-06
Livonor Wrote:I've been learning 400-1000 words a day, if that's not being efficient I don't know what is.That's not a typo is it? If you're seriously "priming" up to 1,000 words a day then that certainly is pretty interesting. My guess is the majority of these words come from predefined lists rather than via mining or encountering them naturally whilst reading? Otherwise that just doesn't seem feasible at all. I'm also wondering how you plan on "retaining" this "knowledge". Are you at a minimum able to recognise the word and recall meaning and pronunciation? If you're not using a SRS, what's your routine like and do you keep track of retention?
EDIT: For reference, when I was at an intermediate level I'd probably average adding 40 words or phrases per day (using Anki) - the most I've added would've been not much more than 100 per day. I'm just struggling to understand how someone could conceivably be at a level where learning to recognise 1,000 words per day would be worth the effort. If you're at a level where adding 1,000 words per day is somehow useful to you then I think it'd be more cost beneficial to focus on depth (i.e. quality) rather than breadth. But that's just my opinion.