So I'll be finished with RTK1 in just a few days and will then be moving on to sentences but I've been playing around with a few ideas as of late on how we could really expand on the method and possibly combine the two main ideas of AJATT which of course is 10,000 sentences + constant immersion (usually via some form of audio).

What got me really thinking about this is the little Shadowing CD and booklet I have. I've been playing around with it from time to time when I'm not working on RTK1 and I came to realize just how little time it takes to acquire and understand a 30-40 second dialogue consisting of roughly 15-20 "sentences"...especially when you can read along with it at first.

I'd honestly say it only takes maybe 10-15 minutes or so before the entire dialogue starts to click and you understand just about everything in it. Basically my method is this:

1. Put the dialogue on repeat and just read along with it silently for the first 2-3 times. This is just to really make sure you know what is being said for any harder dialogue where you can't pick apart the sounds with just your ear.

2. Start shadowing while still reading along for the next 5-6 plays.

3. Put the book down and do the next 5-6 plays with just the audio only. While you're shadowing just really think of it as a conversation and try to imagine what is taking place.

As I said before that takes all of maybe 15 minutes or so and by then I've usually "got" the dialogue pretty damn well.

The kicker or course is that in that 15 minutes I just learned 20 sentences all in the context of real spoken dialogue while ALSO getting to exercise my own speaking ability. Throughout the day just pop in that same dialogue in maybe 10 minute spurts on repeat, alternating between shadowing and just pure listening and by the end of the day you got yourself and entire contextual spoken dialogue pretty well cemented in your head.

Now think about this. If you did only ONE such dialogue per day in one year you would have 7,300 spoken dialogue "sentences" that you would not only understand but could also speak. For those "hard core" among us that number could easily be doubled if not tripled.


So how could this be implemented with our SRS's such as Anki? Well we all know that plenty of people include audio for just their single sentences but for this method that would be a complete waste of time. My idea is that we would create entire MP3 play lists and/or CD's full of our dialogues and our Anki card "question" would simply be a note telling us which playlist/CD track. You also wouldn't need to have just one dialogue per track if you really wanted to. IMO you could easily have several dialogues per track because even a 60 sentence dialogue would only take up up a couple minutes at the most. Doing it this way in MP3 format you could easily have months worth of "cards" even on a single CD.

For example let's say you only do one dialogue per day. Maybe every three days you put those 3 dialogues together as "Track 1" in your MP3 play list. You then add an Anki card to your deck simply labeled "Listen to track 1" and maybe include the written text of the conversation as your "answer".


Any input or suggestions? When I finish up with RTK1 I believe I might give this method a shot and try to work out the details.

It's not that I think this is a bad idea, but putting so much effort into doing this simply isn't necessary. Your goal is to improve your listening comprehension, but your ability to understand will compound over time. Assuming that you keep progressing with your studies, I'm sure that even if you don't listen to your Shadowing CD for another two years, you will be able to listen to it then and have complete comprehension.

It's certainly a good idea to listen to the CD until you understand it, but your time would be better used listening to new material rather than going over audio you can already understand.

samesong Wrote:It's not that I think this is a bad idea, but putting so much effort into doing this simply isn't necessary. Your goal is to improve your listening comprehension, but your ability to understand will compound over time. Assuming that you keep progressing with your studies, I'm sure that even if you don't listen to your Shadowing CD for another two years, you will be able to listen to it then and have complete comprehension.

It's certainly a good idea to listen to the CD until you understand it, but your time would be better used listening to new material rather than going over audio you can already understand.

I'm not really following you. I would be adding new material every day. I'm not talking about just using this one Shadowing CD and putting it on repeat for the next year. I'm talking about taking dialogue sources, such as movies, and breaking them down into smaller bites of dialogue that you then learn via listening and Shadowing for a day.

What we do right now with the current methods is take text based sentences, break them down until we understand them and then hope that via our immersion practices we will begin to start hearing bits and pieces from those individual sentences.

What I'm saying is why not combine the two and skip the waiting around to hear it part? You can get an entire day's worth of sentences in just one minute or so of audio and learn an entire little conversation in the complete context of a whole discussion.....not just the context found in a single sentence like with our current methods. The context from one sentence is great so context from a sentence and it's contextual response would be even better because that is what makes communication.

The part that would take most time would be gathering the audio clips that have translations along with them. If you could somehow cut out that step, then you've really got something going.

danieldesu Wrote:The part that would take most time would be gathering the audio clips that have translations along with them. If you could somehow cut out that step, then you've really got something going.

Yeah that's were the effort will be but I honestly don't think it will be that big of an issue once I figure out a audio recording setup and the best/easiest material to use.

For instance if I just want to learn one 20 sentence dialogue per day that's only going to be a minute or less of actual dialogue. You could sit down one day every couple of weeks and just run through one or two episodes of your favorite Japanese subtitled J-Drama or Anime and have enough dialogues to last you a month or so.

If we got just a few people to help out we could easily amass a huge library of audio dialogues.

Getting the text should be straight forward enough. Simply play a clip and copy the subs. The only real difficulty will be finding the easiest method to rip audio to MP3 format.

It's an excellent Idea - something I've been thinking about how to do too.

Could you get the clips from a Japanese Learning Podcast like japanese101.com.
Do they have English Transcripts too that you could use straight away?

My idea was that you could put the Vocab and Sentences into Anki at the same time for additional learning.

The trouble is that Learning Podcasts aren't "real life speaking". To do that I guess you should take clips from TV shows or Movies. This sounds like it's going to take a lot of time - also, how are you going to find the English translations?

Perhaps this could be set up as a collaborative effort? If you had many people each contribute several "Audio sentences" then together you would be able to build a really awesome learning tool.

In fact if you had 10000 audio clips with translations, it would be awesome for learning Japanese really fast. Perhaps the only tool you'd need!

Well, you'd still need to learn how to read. And besides, shouldn't you go monolingual some point along the line, instead of having 一万 translations?

I'm actually thinking I could simply do this the "old fashion" recording way which would make it exceptionally easy to do. Sometimes I get caught up in trying to do it the technologically "correct" way (whatever that means lol) that you forget about the simple methods.

The way is to just buy a decent quality recording microphone for my computer and simply let it record straight from my DVD player's speaker output. All think I'd have to do is buy a decent speaker and enough cable so I can run it somewhere to isolate the speaker and the microphone, such as my bathroom.

No it wouldn't be very tech savy but it would make the entire process a simple point and click operation. No worrying about DVD ripping, security codes or trying to find Japanese subbed videos online to work with. Pop in any Japanese DVD into my DVD player, press record on my sound player, stop when I'm finished, save file as, done.

Doing it this way I could record an entire TV episode worth of audio dialogue and separate it into 20 sentence bites in as about as much time as it takes me to just watch it by itself.

playadom Wrote:Well, you'd still need to learn how to read. And besides, shouldn't you go monolingual some point along the line, instead of having 一万 translations?

Who says you can't do it monolingual? When I say copy the subs I'm talking about copying the Japanese subtitles, not the English ones. How you translate them after that will be up to you just as it is now with the current sentence method.

activeaero Wrote:

playadom Wrote:Well, you'd still need to learn how to read. And besides, shouldn't you go monolingual some point along the line, instead of having 一万 translations?

Who says you can't do it monolingual? When I say copy the subs I'm talking about copying the Japanese subtitles, not the English ones. How you translate them after that will be up to you just as it is now with the current sentence method.

I suppose I misunderstood you -- I'm assuming you meant transcription instead of translation here -- cause the language is remaining constant.

But yeah, if done with Japanese subs, there's probably a LOT of power in this waiting to be unleashed.

Quote:I suppose I misunderstood you -- I'm assuming you meant transcription instead of translation here -- cause the language is remaining constant.

Yes you're correct. My bad.

Quote:But yeah, if done with Japanese subs, there's probably a LOT of power in this waiting to be unleashed.

That's what I'm thinking. It really makes no sense to keep the audio immersion and sentences separate. Our sentences should be our audio immersion and vice versa. It's hitting the brain from all levels. Instead of just reading the sentence you're learning to understand it as it comes from the mouth's of native speakers, learning how to say it and to top it all off you get an awesome mental association based from whatever movie scene the dialog is from.

IMO it will also be a huge morale booster because you will actually be learning Japanese that you can apply right away every single day.

I mean imagine doing this start to finish with your favorite movies/animes/dramas. Every single day you are learning to understand an entire conversation from something you ENJOY watching......instead of just learning some lone sentence all by itself. For example let's assume your average 25 minute anime episode, or whatever you fancy in terms of shows, has about 10-15 decent dialogues. Well that means in just a week and a half to two weeks of doing just one dialogue a day you now understand an entire episode from start to finish! Since I see no reason a hard working person couldn't double or triple the number of dialogs they do each day I don't see why someone couldn't learn to be able to understand an entire episode from start to finish every couple of days.

Assimil is a language course often used like this, it has 100 short dialogues.

activeaero Wrote:I'm not talking about just using this one Shadowing CD and putting it on repeat for the next year. I'm talking about taking dialogue sources, such as movies, and breaking them down into smaller bites of dialogue that you then learn via listening and Shadowing for a day.

I'll try to clarify my previous post.

I understand you won't be listening to only the Shadowing CD, but if put into your SRS, you will be listening to it repeatedly over the next however many years it pops up in your deck. As long as you keep listening to new material, you won't need repeated listens to understand the CD; listening is something that comes naturally after so many hours of exposure. Using an SRS to improve listening comprehension simply isn't necessary.

It will take hours and hours of your time and effort to compile all of the sound bites. Over the course of the year, you could spend 50-100 hours putting together all the sound bites for your SRS, or you could spend that time watching J-Dramas, listening to a pod-casts, keeping up on the news in Japanese, etc.

It seems like a good idea, but your time could be much better used to improve your Japanese.

I recommend playing around with this website:

http://voice.pentax.jp/

You can enter Japanese text and then you can choose a male or female voice to say it. I was copying and pasting things from Japan Wikipedia and it's pretty good for shadowing or emulation.

Shorter sentences sound a little funny from my point of view, but give it a shot.

I have two visual novels, each about 15 hours long--90% dialogue, 10% narration--and fully featured with audio for the dialogue. If you've never seen one before, here's a screenshot from ef - the latter tale: -link-

A lot of the sentences are still way beyond my understanding since I didn't start sentences too long ago. I never even thought about using them for my SRS. My main reasons for listening and reading along are to build my reading speed, dictation skills, and to pick up readings naturally for kanji.

However, I have thought about putting the dialogue from ef into Speech Analyzer and see how well I can shadow it. I haven't played around with this program much yet, but Stuart Jay Raj recommended it for understanding how sounds are made and comparing yourself.

Visual novels would likely be perfect for this, but I don't know how common it is for them to have audio. Most of the ones I've found are text-only, for example, 「ひぐらしのなく頃に」.

samesong Wrote:I understand you won't be listening to only the Shadowing CD, but if put into your SRS, you will be listening to it repeatedly over the next however many years it pops up in your deck.

Which is the exact same thing we do for sentences now so I'm not exactly sure how that can be a bad thing.

Quote:As long as you keep listening to new material, you won't need repeated listens to understand the CD; listening is something that comes naturally after so many hours of exposure.

Can't the same thing be said about reading sentences?

I mean why is it ok to learn how to speak and understand Japanese by READING text based sentences in our SRS but a waste of time to do the same thing by using the actual REAL dialog that we want to speak and understand?

Quote:It will take hours and hours of your time and effort to compile all of the sound bites.

Actually it shouldn't take much longer than it does to do the current sentence method if I use the speaker/microphone recording method I described above. Making an MP3 file of the audio will be the amount of time it takes me to hit "stop record" and "save file as". Breaking down the sentences will take just as long as it does right now.

Let's take a 25 minute TV episode. Pop in the DVD and as it plays just hit "start record" and "stop record" for every dialogue segment you want. You'll only have to pause the DVD for the brief time it takes for you to make the saved file name for each clip. The total time to do one episode shouldn't be much longer than 30-35 minutes and now you've got 10, maybe 15 dialogues. Now go back and watch it again and grab the sentences from the subtitles just like you would with the current sentence method. You've now got dialog audio and transcripts for an entire week or two of studying and all it cost you was maybe 2-4hrs of "work". Now simply do a dialogue each day and you can add the transcript section to ANKI just like you do with individual sentences right now.

Not to mention the whole "work" process of gathering the dialogues just doubles as immersion time anyway so I don't really see much of a downside to it.

This is kind of what is being done, while watching dramas with the script/japanese subs. What was that... rewind...

One thing though, you might want to shadow a member of the same sex, or so I've been told.

activeaero Wrote:I mean why is it ok to learn how to speak and understand Japanese by READING text based sentences in our SRS but a waste of time to do the same thing by using the actual REAL dialog that we want to speak and understand?

Again, it's just a matter of how your time is spent. You can get through 200 reviews in Anki in a pretty short amount of time because all that is required is to quickly scan the sentence and make sure you still remember/understand the content. Down the road, the time spent reviewing all of your clips is going to be MASSIVE. A dialog might only be between 10 and 30 seconds, but doing 50 to 100 of those a day (depending on how many reviews you have, of course) is going to be very time consuming.

But I'll try to emphasize my earlier point: re-listening to dialogs just isn't necessary one you have comprehension of the material. We use an SRS to hopefully pound vocabulary and grammatical structures into our brain (which we already have an understanding of) so we can recall them at a later time. You can gain the same understanding through simply reading the sentence, which is much faster. There isn't anything specific or special in an audio dialog that you couldn't get out of text. Correct intonation and pronunciation, listening comprehension, and speaking ablities are gained by massive amounts of exposure to the language, and if you are listening to the same clips over and over again, the level of exposure is minimal.

Anyway, good luck with the project. I hope it does work for you.

Quote:Actually it shouldn't take much longer than it does to do the current sentence method if I use the speaker/microphone recording method I described above.

Another subject entirely, but I do have an issue with how much time is spent mindlessly mining rather than studying/reading/listening, etc.

I do have a question. Where did you get your visual novels?

If it matters, iKnow website lets you get podcasts of your lessons from the last 20 days. Each 10 item lesson is about 90 seconds in length (it says the vocabulary word, then the sentence using that vocabulary word). Just download into your iPod, and listen till you're board for the day.

That should save you loads of effort.

That said, the immersion is listening to real Japanese. That can be movies, dramas, radio podcasts, anime, etc. It's not about listening to one thing till you get 100% comprehension. It's listening to everything, with attention or even being distracted, all the time. As your vocabulary and grammar increase (slowly but surely), you start comprehending more of what you're hearing (slowly but surely).

samesong Wrote:Again, it's just a matter of how your time is spent. You can get through 200 reviews in Anki in a pretty short amount of time because all that is required is to quickly scan the sentence and make sure you still remember/understand the content. Down the road, the time spent reviewing all of your clips is going to be MASSIVE. A dialog might only be between 10 and 30 seconds, but doing 50 to 100 of those a day (depending on how many reviews you have, of course) is going to be very time consuming.

You wouldn't have 50-100 reviews a day. Remember a 1 minute dialog might contain 20+ sentences. If you added two dialog's a day you're only ever going to have reviews of 4-5 dialogs per day at the most which will take you all of 5-10 minutes to listen to.

Quote:But I'll try to emphasize my earlier point: re-listening to dialogs just isn't necessary one you have comprehension of the material.

Says who? I could say the same thing about sentences: "oh once you learn it the first time you'll get it every other time so there is no point putting sentences in an SRS".

Quote:We use an SRS to hopefully pound vocabulary and grammatical structures into our brain (which we already have an understanding of) so we can recall them at a later time.

And that's exactly what this will be doing so I don't see your point.

Quote:Correct intonation and pronunciation, listening comprehension, and speaking ablities are gained by massive amounts of exposure to the language, and if you are listening to the same clips over and over again, the level of exposure is minimal.

The same thing can be said for reading the same sentences over and over again.

Maybe you're confused and think I'm saying to use these dialogs as your ONLY source of input. I'm not. You would treat it just like you do now with sentences. You'd still listen and watch as much material as you could outside of the dialogs you are learning. The point is the exact same as sentences. As you become familiar hearing and understanding all of what is being said in your dialogs you'll start to understand more of the material from other listening sources.

activeaero Wrote:So how could this be implemented with our SRS's such as Anki? Well we all know that plenty of people include audio for just their single sentences but for this method that would be a complete waste of time. My idea is that we would create entire MP3 play lists and/or CD's full of our dialogues and our Anki card "question" would simply be a note telling us which playlist/CD track. You also wouldn't need to have just one dialogue per track if you really wanted to. IMO you could easily have several dialogues per track because even a 60 sentence dialogue would only take up up a couple minutes at the most. Doing it this way in MP3 format you could easily have months worth of "cards" even on a single CD.

I'm of the opinion that single-sentence items are probably the upper limit of what should go into an SRS. The minimum information principle is one of the golden rules of SRSing, after all.

That's not to say that listening to dialogs or watching movies is a bad thing, just that throwing things into Anki isn't always the best solution.

activeaero Wrote:

Quote:But I'll try to emphasize my earlier point: re-listening to dialogs just isn't necessary one you have comprehension of the material.

Says who? I could say the same thing about sentences: "oh once you learn it the first time you'll get it every other time so there is no point putting sentences in an SRS".

Quote:We use an SRS to hopefully pound vocabulary and grammatical structures into our brain (which we already have an understanding of) so we can recall them at a later time.

And that's exactly what this will be doing so I don't see your point.

'Cause, there's a difference.

You can't learn a word or kanji reading once and never forget it. Learning to comprehend audio, on the other hand, is a built up skill that isn't something you review in increasing intervals. You don't "learn" to comprehend specific dialogs with a finite amount needed for listening comprehension.

Plus, I mean, real life spoken Japanese is a lot different than learner's dialogs give. I don't think learning a lot of dialogs from learner's sources like this will necessarily compliment your ability to listen to real Japanese any more than normal AJATT.

activeaero Wrote:stuff

Awesomeness. Give it a try; maybe I'm wrong

Maybe you missed my previous question: where did you get your visual novels from? I looked into it and it seems like a pretty fun way to learn Japanese.

The way you've described listening/reading the dialogue, shadowing the dialogue while reading, and then shadowing the dialogue while not reading sounds JUST like the method used in many places (everywhere?) in Japan. Trust me. I do it every day. EVERY DAY. *cry*

Anyway, if you've been to Japan, you will know that most people who go through this system are great at English if someone drops their change (Excuse me. Is this your change? Oh, my change. Thank you. You're welcome.) but can't answer basic questions like "Where are you going?". What's great is they can use this conversation anytime change comes up in the conversation or is seen or is in their pockets or within a 3 kilometer radius, whether it's applicable or not. If you want to speak like that, then do this audio dialogue method as you've described it. But by just altering what you've described a little it can be very effective in making you a native-like speaker.

As far as samesong saying it's not much use listening to something you've already been able to understand, that's not true. In fact, you are only acquiring vocabulary/grammar when you understand what you are hearing/listening to. (That's if you agree with the Input Hypothesis upon which AJATT is based.) For me, moving on to news, radio show, book, manga, etc I've never heard/read before means lots of stuff I don't understand and must struggle to understand. Still it's what I do when I want to approach something new. When I'm reading I take the time to look things up and work meaning out. When I watch things or listen, I just live in a state of semi-understanding.

So, listening to these dialogues a few times and figuring out what they say is normal. And once you understand them, listening to them again from time to time is a great idea. It's probably more helpful to listen to a dialogue you know you'll understand 2 or 3 extra times than it is to listen to 2 or 3 new dialogues that you know you'll struggle to understand. This is in fact the whole idea behind AJATT's SRSing sentences--you'll have more exposures to sentences you know you'll understand but won't have the opportunity to memorize (assuming you are inputting enough). As you've pointed out, Activeaero, using these dialogues allows one to hear an even bigger context than the sentences do. Depending on how you SRS with sentences, including dialogues to be heard could be a nice break from the sentences or it could be an interruptive factor--possibly an activity better given it's own set-aside time or separate SRS deck.

In general memorizing dialogues (which is how you've characterized this approach) helps no one become a native-like speaker. As far as improving your speaking, input, as samesong said, is the best way. This is important. So again, input--listening and reading--is the best way to improve your speaking. Big fat period. Doing the shadowing while reading is a conscious exercise in pronunciation. Not a bad thing in and of itself. But it's akin to putting icing on the cake. Doing shadowing while not reading is basically an exercise in memory and not helpful in terms of language acquisition. It can be detrimental due to our very imperfect memories.

But in general the idea of re-listening to spoken language (just like re-reading text) you've already decoded is a good one. It's better to spread out the listening so that you're not relying so much on your memory of the dialogue and more on your linguistic competence. Also, I'd like to agree with samesong's criticism of sentence-mining. The biggest point is living in Japanese. So you should only take the time adding dialogues (or other spoken language) that you found interesting for whatever reason as you were listening to it in your daily life.

I'd say the main problem with this method (outside of how much more effort there is in this compared to the normal sentence mthod and I really doubt the payoff is worth it), you will only learn conversational Japanese since you won't find anything else in dialogue form. If you want to be able to read essays, novels etc, it might not be a good idea.