I dont think its a good idea, because the sentences arent originating from native speakers. So, they might be 'correct' per se, but still perhaps a bit odd.
Of course, if someone really wanted to do that, they could just look at many of the corrected entries that are already viewable on there...

Of course I can find sentences. I can't find them with readings and translations that I can easily put into a SRS. I can't just ram my phone bill into the DVD drive of my computer.
I know they aren't coming from native speakers, but this could be fixed by having a native speaker look over them and make sure they don't sound too strange. Besides textbooks aren't exactly natural and even native speakers don't all speak the same way, so even when you are learning something from a friend you might be picking up something that only your friend says or only certain people say. I don't think you can "sentence mine" that problem away. The only way to learn what words people say most often is to speak to Japanese people and read lots of Japanese from different sources. Unless you're planning to do 100,000 sentences I don't know if you can learn to speak and write naturally through SRS(please don't say AJATT, because I've had people tell me his Japanese is good, but not like a typical Japanese person).
My idea was just a way for people to get some free sentences, so they didn't have to spent there time preparing to learn Japanese. If I have 2 hours to study everyday, I want to spend at least 1 and a half of those hours studying. I don't want to spend 1 hour looking for(typing) sentences.

Chandlerhimself wrote:

I can't find them with readings and translations that I can easily put into a SRS. I can't just ram my phone bill into the DVD drive of my computer.

I see. I guess dictionaries are out of fashion these days.

Its not a bad idea, but I think one of the strengths of using a book like KO is you are making the most use of your study time, in the sense that you are studying the Highest frequesncy kanji compounds and voc.

After RTK, I did a little sentnece minning, and surfing websites for sentences, but I found it too time consuming and also I wasnt ever sure if  word I found was a common word or not, without checking.

The sthing about KO is it prioritizes your Kanji compunds, readings, and voc study, so that you can be up and reading most things (60-70% at book one, 90% at book two) asap.

I felt that if a lot of people were contributing to blogs, then you would get a range of sentneces for sure, but they might range from Final Fanasty words to computer terms, and while that may be a good thing in the long run, I think compound frequency study is a better plan of attack.

Still, the idea has some merit. Perhapase we could make a KO book 3 with the frequesncy list for the kanji that KO 1and 2 dont cover, using lang 8.

Lang-8 has a lot of Japanese written by native speakers as well.  I wouldn't enter in corrected sentences too much, since they started as non-native and then often they are corrected to be passable Japanese, but maybe not the best.

Lot's of Japanese people use the site and write comments, blog posts, and forum topics in native Japanese.  So those would probably be a good source. Just bear in mind the style, which is written Japanese for the internet...which will of course be different than spoken Japanese or more formal writing.

Chandlerhimself wrote:

zazen666- I don't think this should replace KO I think you could use both at the same time. You said when you started it took too much time to mine for sentences. I just want to give people like you(and me), who don't have time, something they can put into their SRS and then use their time for studying and not searching. Your idea for a KO vol3 is good. I think it would be great if everyone got together and made a list of sentences using those kanji.

For me it makes no sense for everyone to make their own cards when we can all just share cards. I'm sure everyone on here has different things they want to do when they learn Japanese, but for the most part it wouldn't be a bad idea to use another persons' sentences if it saves time.

I like your idea in theory-the KO project did save me a LOT of time (i typed in half the book myself before the project was born). Hmm, maybe we should start a new thread, since this one seems to have jumped off track. (I dont people should jump down your throat for throwing out ideas)

We could consider the KO vol3. get the list and divide up the Kanji. It doesnt necessary have to use lang 8. We could track down sentences and submit them.

Yet another idea I had was to create SUbject based sentences. I am reading some books on Astronomy now in Japanese and I search around on the internet a bit trying to find a decent list of related terms to plug into Anki to give me a heads up. Words like, solar system, galaxy, light year, light speed, oxygen, hydrogen, shooting star, etc etc.

I thought something like that might be usefull. Loke 50 sentneces with specific voc per each topic. Topics could be what ever. Science, SF, History, Music, Games, whatever.

for me persoanlly, anything like this would really be one the side for me. It took me a year to get thru KO1 and I am satisfied with the results, so I plan to keep on trucking thru KO2-which might take another year-and just read books for fun on the side.  Its really had to say what the needs will be a year from now.

Last edited by zazen666 (2008 September 03, 4:29 am)

Chandlerhimself wrote:

I don't think you can "sentence mine" that problem away. The only way to learn what words people say most often is to speak to Japanese people and read lots of Japanese from different sources. Unless you're planning to do 100,000 sentences I don't know if you can learn to speak and write naturally through SRS(please don't say AJATT, because I've had people tell me his Japanese is good, but not like a typical Japanese person).

You complain that AJATT guy doesn't have natural Japanese, so your solution is to mine sentences that were originally created by non-native speakers?

But like you said right there, the way to really learn is by talking and reading. Now, when you are reading something, you will have to figure out the readings and meaning of the sentences anyways. So this is NOT an extra step that you have to take when adding stuff to your SRS!

And really, I don't see how you can have a shortage of sentences, there are a few other spreadsheet projects going on besides KO, and all of the books in these projects are fairly cheap. There is also the tanaka corpus and the tanuki corpus, where you can find thousands of sentence pairs. And then, there is also the eijiro dictionary ( http://www.alc.co.jp/ ) which has thousands upon thousands of sentence pairs.

So really, there is absolutely no shortage of sentences out there, which are ready to bring into an SRS. I do think there is some merit to collecting and organizing them in a particular way though, such as a "book 3" to KO.

zazen666 wrote:

Yet another idea I had was to create SUbject based sentences. I am reading some books on Astronomy now in Japanese and I search around on the internet a bit trying to find a decent list of related terms to plug into Anki to give me a heads up. Words like, solar system, galaxy, light year, light speed, oxygen, hydrogen, shooting star, etc etc.

I thought something like that might be usefull. Loke 50 sentneces with specific voc per each topic. Topics could be what ever. Science, SF, History, Music, Games, whatever.

The ALC site has a selection of theme-based sentences. There are 50 themes (including "宇宙") and nearly 3,000 sentences in all. It takes time to copy & paste them individually, but if you want them all neatly in a spreadsheet then email me.

In fact those theme databases are only a small part of ALC's sentence collections. The main page is here.

Ah, they were on the old failed AJATT forum for a while, Pauline C & P ed them.  The not yet failed new AJATT forum can be found here:

http://ajatt.pseudosphere.net/

And soon it will also be filled with sentences hopefully.  Dorama, anime and movie scripts alone, if gathered together in one place, would be more than enough sentences for anyone.

Last edited by phauna (2008 September 03, 5:13 pm)

Katsuo wrote:

zazen666 wrote:

Yet another idea I had was to create SUbject based sentences. I am reading some books on Astronomy now in Japanese and I search around on the internet a bit trying to find a decent list of related terms to plug into Anki to give me a heads up. Words like, solar system, galaxy, light year, light speed, oxygen, hydrogen, shooting star, etc etc.

I thought something like that might be usefull. Loke 50 sentneces with specific voc per each topic. Topics could be what ever. Science, SF, History, Music, Games, whatever.

The ALC site has a selection of theme-based sentences. There are 50 themes (including "宇宙") and nearly 3,000 sentences in all. It takes time to copy & paste them individually, but if you want them all neatly in a spreadsheet then email me.

In fact those theme databases are only a small part of ALC's sentence collections. The main page is here.

Hey-thanks for that.

Okay, probalby an absolute newb question, but .... about those spreadsheets in general (I've never heard about them before I came to this site anyway).
How exactly can I put them into Anki?
Copy and paste every single sentence? That's not that much faster than typing, so I suppose there's a faster way to do it?
Please enlighten me.

Thanks and sorry for the OT.

phauna wrote:

I agree that the ideal is to never, ever type up any sentences, just use all that time to post links to forums, ah..., I mean study more.  So I proposed before to have people sign up for a free blogger blog and post some non-copywrite stuff that they find.  If everyone did it then we soon would have a large deposit of authentic Japanese which would forever be available.  Anyway, I talk about it in this thread:

http://forum.koohii.com/viewtopic.php?id=1831

On the subject of copyright and subject-based collections, wouldn't just taking a few sentences from published books be fine? Not whole paragraphs or anything like that, just a couple of choice sentences? Does that fall into 'fair use'?

Using the books about space as example and blogs as the method of delivery, you could take sentences with the more important words (star, planet, comet, solar system, etc.) and post those. Or start a blog for a novel/manga/drama/etc., and take 5/10/15/however-many-you-like sentences you find interesting or useful from a chapter or episode to post.

The goal wouldn't be to have a complete catalogue of every sentence and phrase, just a couple of examples. By keeping a limit on how many you take from each chapter/episode (which you can go beyond if you want), it doesn't seem as massive a task as "mine the whole of 涼宮ハルヒの憂鬱" and hopefully makes it easier for people to get into and keep going.