http://www.sciencentral.com/articles/vie...=218393117

This article about a new study on 'tip of the tongue' remembering advises not to strain too hard trying to remember something. This will only cause you to always require a lengthy working out period to remember it. Rather you should look it up immediately if the answer doesn't come to you quickly.

This reminds me of Khatzu's advice not to 'try' to remember your SRS sentences. Just go through them, pass or fail them, with time you'll understand easily. I usually use this method more through laziness, I just can't be bothered trying to come up with a long string of relationships to get to the answer.

I admit, I do this as well with the study program I use out of laziness, mainly. If I don't remember it immediately, I take a fail, slap myself for forgetting (figuratively), and look at the answer.

Some words I've had to do that 10-15 times before they sunk in, but now they are stuck REALLY good. Better than many of the ones I found easy from start, in fact.

It's nice to see that my laziness is paying off for once.

Cool post, phauna. Like wccrawford, I thought I was just being lazy on this point. Now I can say I'm being more efficient.

I'm with you guys.
Tks phauna for making me not feel bad about beeing lazy.

Thanks! I was a bit torn myself between letting go or trying to remember even if it takes a long time.

I heard someone on this board talking about some research that suggests that memories which are deep in your long term memory take longer to dredge up. I hope someone will link that study up for us. I suppose there is some optimal amount of time to produce a fact from memory, not too long and not too short.

I have found myself sometimes getting stuck on a long, contrived method to remember a word as I've been going through the KO sentences. Something like remembering another word I know well which contains the kanji for the word I want to remember, then trying to guess the sound and lastly trying to nut out the meaning. So I'm going to try in future just to fail it and see the kana and meaning quite quickly after.

Ahhhh great article. I feel it explains nicely why I repeatedly get the same kanji wrong. Fail quickly and break the cycle of the wrong neural path, creating a new and correct one.

Yeah, I'm not fond of the anti practice, anti class rhetoric of khatzu and antimoon. Maybe you will damage your Japanese if you even 'think' about a mistake? Maybe you shouldn't think about Japanese at all, just in case. It's kind of ridiculous. Fluency is as important as accuracy. They more you study the more you can work out the kinks one by one. I can't wait ten years to have perfect Japanese, I'm living in Japan right now. I have to talk to the konbini guy and the taxi driver.

I'd really like to hear Khatzu and Antimoon guys use their language, I think you will not be so in awe if you did. They will probably sound like guys who study hard, like other people you've met who went to classes and did grammar exercises and never used an SRS.

These people look at mistakes in the wrong way. They think they are black and white. Imagine not saying anything until you had perfect pronunciation. With pronunciation it is easy to see there is a continuum from terrible to perfect. How will you improve this without using it? Input is important, and maybe a large amount of input at first will give you a good overview of the language. However you have to say something sometime, speaking is the most important skill by far.

Also, classes are useful if they are good. Meaning they have a lot of structure practice, but not much wordy explication by the teacher. Good language learners are risk takers, not careful nellys.

phauna Wrote:Good language learners are risk takers, not careful nellys.

Stephen Krashen says there is a balance between risk taking and being a careful nelly. Of course the risk takers end up being the ones who can rattle off in a conversation at native speed, but they make lots of mistakes a native would not make. The careful nellys may always use simple language, not feel confident, and speak very slow, which is also not good.

I think I agree with you though, phauna. The way the brain works is different between inputting and outputting language. I feel like I learn a pattern better if I have used it. How can I use it if I haven't learned it? Obviously I have seen the pattern in my input, so I understand it. So when I need to say something, and I look up in Google how to say it and I see that particular pattern being used, then I make that connection. But my brain will not automatically connect the input to the output except for basic patterns and words.

What antimoon is saying, and I think what katsumoto doesn't get across very well at AJATT, is that as soon as you want, you can start speaking in basic sentences. This could theoretically come very soon. In fact, I think the more you do this, the more you are used to using the language, the sooner you will get tired of using just the basic stuff and you will see more easily how to fit the more complicated patterns into your speech.

The antimoon guys have voice samples on their website. Not what I'd call flawless, but more impressive than most I hear.

The avoiding mistakes thing has always made a lot of sense to me. If you get enough of the correct input (and imitate accurately), it'll start to come out right in the first place, and you won't need to practice doing it wrong before you get it right. You don't practice getting the right form on the piano by putting your fingers flat and then working up to a bubble. You start with the bubble, or else you spend a lot of time relearning things later when you realize the way you were doing it just won't cut it anymore, or else you never become a competent piano player. The same is very transparently true of any other physical task (which pronunciation is) that you can think of; start correct, stay correct. It's true of mental tasks as well, but more difficult to express convincingly.

More than anything, I find the idea convincing because I've seen enough people who learn by output and (consequently) mistakes first. People who come here and live their whole life here and still make the same mistakes over and over. Or people who receive incorrect input; many of my relatives and particularly my father, who has been speaking in an uneducated manner for so long it's pretty much impossible for him to stop now.

"Fluency is as important as accuracy" isn't a statement that makes a lot of sense to me, as fluency, while measuring ease, implies that the ease involves accuracy. I know people who quite easily speak 'English', but 70% of the time I can't understand what the hell they're trying to say because the English that comes easily is also a grammatical train wreck and the pronunciation isn't any better.. I'm not inclined to call that fluency.

I'm not sure where you've seen a progression of terrible to perfect as related to pronunciation. I've never encountered this myself. Normally I see a pronunciation of bad progress to bad, or good progress to... good. This comes from first-hand experience and watching, for instance, celebrities as they learn the language. An example that comes to mind very quickly is Jackie Chan, who had to start reciting English lines before he knew any English at all, and now, at this point, is still difficult to understand in a movie and incomprehensible outside of one. But I'm not saying my experience is all that broad, or that it's impossible - you CAN always improve something that you're doing wrong. I'm just not sure it's a reliable trend to hope for.

It's not that you should never speak, either. Not that you should never write. It's fine to speak while you're imitating someone else as accurately as possible. It's not alright to speak with no basis for comparison, until you've already got a grasp on whatever it is you're trying to say. It's fine to write out native speaker's sentences. It's not alright to write your own sentences until you're already sure what you're writing is correct.
I think this actually goes along with your initial post, which has me confused. As you said, you don't want the wrong path to build there. If you're sitting there struggling for a kanji story, you're using and building on the wrong pathway; get rid of it and make the correct one. The same is true of everything else; don't practice speaking incorrectly, or you're only building and reinforcing the wrong mental paths. Just imitate a native speaker and build the right ones.

As it happens, I think I'm the one who said something about deep long term memories being harder to withdraw (http://www.wired.com/medtech/health/maga...ntPage=all). I think that's still true, but as relates to kanji recall I think I've changed my mind on the relevance. Deeply ingrained somewhere or not, renewing the correct path for active recall seems like a good idea.

If you are actively trying to improve, I think you can constantly improve. All those immigrants who move to English speaking countries may quickly decide they have enough to get by and forget about studying ever again. These are not language learners, really, though they are used by Antimoon as the example of bad accuracy/ good fluency learners.

I think lots of people see this passive propaganda and think that if they get enough, one day they will instantly be able to say what they want. I think massive passive exposure is necessary, and some active use at modest levels also. Of course you should be able to understand more than you can produce, it's the same in a native speaker, we all probably know heaps of words which we never say day to day, and may never have ever used.

'Myrmidon' is an example of a word which I have never said aloud nor written down, only passively encountered in reading.

As for pronunciation, what is good, and what is not? There is no answer, really, only intelligibility. Even sounding like a native speaker is not such a good measure, what kind of native speaker? So there are many degrees of ability, the same as with other skills. Maybe the speaker can't use a Japanese 'r' but can make other sounds, maybe 'fu' is difficult for them, or easy. Americans couldn't pronounce themselves out of a wet paper bag, to my Australian ears, however does it make their pronunciation worse in a language sense??

phauna Wrote:Americans couldn't pronounce themselves out of a wet paper bag, to my Australian ears, however does it make their pronunciation worse in a language sense??

It sure does make a difference where you're brought up! I'm an American that's lived in Australia for the last year and a half (and moving to Japan in a few months). I think many Australians have horrendous pronunciation - especially men- all the words are run together "eymateyawannagodowntotheservongrabbrekky." There are also a lot of Australian only words (brekky, arvo, etc) and borrowed words from England that Americans don't use (pissed-as in drunk, servo-gas station, boot-trunk, etc). Most foreigners learn American English because they've either been taught it or they pick it up through movies and TV- on top of the accent it makes Australians extremely difficult for foreigners to understand. I even had problems the first couple months I was here (once my vocab expanded it was much easier- I bartended here for a while and found out there are also sorts of nicknames for beers/liquors as well)

I suppose where the American is from makes a huge difference as well- those from the south can be just as hard to understand as Australians. I'm from Chicago and foreigners are relieved to talk to me here in Australia because they can actually understand me.

Sorry for the tangent, but since I have so much experience with it I thought I'd comment- such an interesting topic. Is it the same in Japan with the different dialects?

edit: I should add- I prefer the Australian accent now- hearing a US accent is... borderline annoying I also thought of another thing that makes Australians harder to understand for foreigners- different pronunciations of the same words (oregano, zebra, alumin(i)um, etc)

So, the thing is we need to figure out what is best when reviewing. If we know right away the kanji, no problem. If we don't know it right away, should we try to remember?, making associations, comparing in our mind with other kanjis, etc or just mark it failed and move on?
Until a couple of days ago, I took a lot of time trying to remember difficult kanjis, but after reading this thread, I started to accelerate my reviewing, decreasing a lot my retention rate (from 90% to 60%). I don't know if it will pay off in the long term.

I think if you know the answer, but not right away, you should think about it. Won't it be strengthening the connections?

yukamina Wrote:I think if you know the answer, but not right away, you should think about it. Won't it be strengthening the connections?

The point of the article is that thinking about it hard DOES strengthen connections, but if you are struggling with something, you might be thinking hard and then come up at the wrong conclusion. Then, you have just successfully strengthened incorrect data into your memory!

Interesting....

That's actually the complete opposite of what I instinctively thought. I always thought trying to remember something had a positive affect, but that study makes a lot of sense. I will try it!

[Katsuo]

If you view memories as networks of connections, then maybe when a direct link is weak, broken or blocked somehow, the memory can still be located (eventually) by going on a series of detours.

If that is the case, then it's important not to reinforce the circuitous route. It would be better to somehow repair the direct link or to create a new direct link.

So in practice the best solution might be:

-- If you get stuck, don't spend too long trying to remember, instead you should quickly review your story (and thereby reinforce the direct link).

-- If you later get stuck on the same story again, then it may need to be reworked. (i.e. the direct link is inherently weak, so you need to create a new one).

Woah, wait a minute here fellas. The original article is not telling us to fail anything we can't answer correctly immediately. Doing so is going to give you a low recall success rate and many more reviews, wasting your time. The only conclusion made in that article is

Quote:The finding: people who had twenty seconds longer to endure the TOT state were more likely to get stuck again on the second test.

Ok so what? Who said TOT is bad anyway? Sure, we don't want TOT when we're trying to use the information, but while we're learning it, as an intermediate stage, it's still better than a 'don't know' answer (and the extra time invested in more reviews). One may even posit that the perfect SRS schedule yields many TOT states.

From the article alone, I'm completely unconvinced that I shouldn't spend time trying to recall. However, let's do some digging because this is quite interesting.

I obtained the original journal article and did one read through of the 7 pages. My conclusion? It's hard to say. First of all, they completely ignore any TOT's that were 'invalid' (i.e., you have a TOT and come up with the wrong answer or when shown the correct answer, it's not what you were expecting.) It's pretty obvious though that these invalid TOTs are no good, since you're reinforcing an error. Second, the results aren't quite presented in a way that is immediately applicable to what we should do as learners. They're more concerned about showing that TOTing once makes you more likely to TOT in the future. Also, they lump together TOTs that are self resolved and those that are not resolved after the 10 or 30 s delay after which the subject is presented the correct answer. Third, increased TOTs on the next review may be a sign of improvement, since TOT is on a level between 'know' and 'don't know'.

I need more time to reread the article to make any conclusions. If you can't access the article, pm me your email address if you want a copy.

___________

What is this confusion about "deep long term memories being harder to withdraw"? That's clearly not the case: how hard of a time do you have recalling your name, address, phone number, etc. etc. It seems the statement you guys are referring to is from the wired article on supermemo linked above. It's a long article, so here is the relevant part:

Quote:One of the problems is that the amount of storage strength you gain from practice is inversely correlated with the current retrieval strength. In other words, the harder you have to work to get the right answer, the more the answer is sealed in memory. Precisely those things that seem to signal we're learning well - easy performance on drills, fluency during a lesson, even the subjective feeling that we know something - are misleading when it comes to predicting whether we will remember it in the future.

Edit: Okay this does make sense. The sense it makes is the opposite of what I thought you guys are saying. In fact, it seems to imply that if we want to increase our storage strength, we should allow ourselves to self recover from TOTs, since these have the lowest retrieval strength.

People could test this theory for themselves to see what works better.

Methodology: Create two decks in anki. Add half of the new material that you learn into one deck, and half of the material into the other deck. Be sure that the material being added is of the same difficulty. Once about 100 items are added to each deck, that should be a large enough sample size. In one deck, never struggle for an answer longer than a few seconds. In the other deck, struggle as long as you want.

About 1 month in, start looking at various stats in each deck. How many reviews do you have daily? How long do your reviews take to complete? What is your retention rate? It would be interesting to see all of this data over a period of about 6 months or so.

Zarxrax: that would be awesome actually. The biggest problem is trying to make sure the material is the same difficulty. That's a really hard thing to do, I think. Any suggestions? I don't suppose anyone wants to memorize stuff like in Ebbinghaus' study: "bes dek fel gup huf jeik mek..."

Not so hard really. Let's say you have a list of 10 new items you want to learn today. Even numbers items go in one deck, odd numbered items go in the other :p

I suppose once you get many items, the random deviations in difficulty should be negligible. It's just that some words from your list happen to be memorable, or you've heard em before, or their kanji are easy, or the kanji are logically related to the meaning; while others are incredibly stubborn. Hopefully it'd even out though.

I'll give it a shot; anyone else?

I guess I might give it a try too, once I start doing sentences in about two weeks.

phuana Wrote:http://www.sciencentral.com/articles/vi ? =218393117

Haha, I go to the university where this study was done