Sentence Mining Explained
I’ve mentioned before briefly of my current Japanese study method, sentence mining, which I first learned about from AllJapaneseAllTheTime. So far it has been working out really well, my reading comprehension in particular has skyrocketed since beginning the method. However, when starting out, I had to more or less figure out on my own how to go about using this method. At AJATT, Khatz explains it in the following steps (I’m paraphrasing):
- Find a sentence with a word(s) that you don’t understand.
- Put that sentence into an SRS flashcard program.
- Find a dictionary definition for what you don’t know and put it in the flashcard.
- Don’t translate, simply try to understand.
This certainly works as a starting point, but it took me a good couple hundred cards before I started hitting my stride as far as sentence mining goes. Now that I’m approaching 1000 sentences mined, I thought I’d take a moment and write down step-by-step what I’ve found works for me when it comes to using this great language study method.
Firstly, this basis of this method is the SRS, or Spaced Repetition System. It is a kind of software that allows you to create flashcards, but it goes way beyond that. By gauging how easily you remember an item, the software employs an algorithm to “space” the next “repetition” of the flashcard. By focusing more time on what you find difficult versus what you find easy, this software simultaneously stimulates long-term memory and makes efficient use of your time.
There are many kinds of SRS software out there, but if you are studying Japanese there is only one choice as far as I’m concerned: Anki. Developed by an Aussie student of Japanese who also knows a bit of code, Anki has become a large open-source project, with many plugins and stat-tracking abilities that make keeping track of your progress a snap. But most important is the Japanese plugin, which makes Anki the best program with which to sentence mine in Japanese. Once downloaded and changing the input method to “Japanese,” the program gives you three input fields: “Expression,” “Meaning,” and “Reading.” Only Expression and Meaning have to be filled in, since Anki automatically generates the reading of your sentence with hiragana. This is a huge time saver that makes using any other SRS pale in comparison.
So now for the step-by-step instructions. First you need to get Anki installed and set up. I’m using Mac OS X, but the directions shouldn’t be much different for Windows. Go to the Anki website and download the version of Anki for your operating system. Get it installed and then launch it, where you should see a window along these lines:
Now you need to get the Japanese plugin installed. Go to File -> Download -> Shared Plugin:
You should now see the following window:
Would you look at that? The Japanese plugin is number one on the list. If you don’t see it, do a search for “Japanese Support” in the search box the top. Once you have “Japanese Support” selected, click “OK” and start the download. When it finishes, restart Anki. Now it’s time to create your deck. When Anki launches again, click “Add Material.” You should now see a window like this:
Currently, the “Basic” model is selected, but we need it to be Japanese. Click on the little wrench icon in the upper-right:
We only have “Basic” available as a model right now, but we can change that by clicking on “Add” and then on “Add:Japanese”:
Hit “OK,” make sure “Japanese” is selected as your model, and then hit “Close.” Now your “Add Items” window should look like this:
You are now ready to start inputting your Japanese sentences! Notice in the upper-right there is a button that says “Recognition.” This is important because you will only be going from the sentence to the definition, never the other way around.
But we need some sentences, don’t we? This part is completely up to you. Whether you want to find sentences from manga, games, books, movies, lyrics, etc., it’s all good. I will say one thing that will make things easier in the beginning: try to mine sentences from websites. The primary reason for this is that it is much faster to copy-paste text from a Japanese website into Anki than it is to, say, find a sentence in a book and then have to look up all the kanji readings so you can type it in. A great tool to this end is the Rikai-chan plugin for Firefox, a clever little Japanese dictionary that looks up definitions based on the text that you are currently mousing over (Pro Tip: use the shift key while mousing over to cycle between the “word,” “kanji,” and “names” dictionaries). With it you can quickly look up readings and definitions while you are browsing Japanese websites and are still getting used to the language.
One thing I love to read online in Japanese is the news. FNN News and TBS News are great, because in addition to the articles there is usually a video with a newscaster reading the text of the article verbatim. You get reading and listening practice, 一石二鳥 (if you can’t read this, why don’t you try out your newly-downloaded Rikai-chan?). Let’s find a nice, juicy article to get some sentences from. Hmm, this looks good:
It’s about an apparent terrorist attack in Indonesia, but there is a word I don’t understand in the title, so it’s a good starting point. First things first, lets copy-paste that title into Anki:
Be sure to paste your sentence into the “Expression” field. After you tab down to the “Meaning” field, Anki automatically generates the readings for the sentence in the “Reading” field. I find that Anki’s readings are right about 80-90% of the time, so if you are just starting out, make sure you cross-check each kanji’s reading just to be safe.
Now lets look up that word I didn’t know, 地元. In the beginning, you will be best served by using a Japanese-English dictionary. But after you’ve been studying for a while, you will want to switch over to a Japanese-Japanese dictionary. There are subtleties in certain words that just aren’t done justice by a simple translation. I like to use the Yahoo! Japanese dictionary. Let’s go there now and find out what 地元 means:
According to definition one, it is the location that has a direct relation to something. In the second definition, it is the location where a person lives, or the location which falls under one’s influence. Coupled with 警察, which means police, I would take 地元警察 to mean the local cops. Let’s grab both definitions and slap them into the “Meaning” section of our new flashcard:
As you can see, I write my definitions by first writing the unknown word, then an equals sign, then the pasted definition. If I have a sentence with multiple unknown words, I break up the definitions with a semi-colon. Once the card is all set, click the “Add” button.
So how do you study these things? Once you are ready to start, click the “Close” button in the “Add Items” window and then begin your review by clicking on “Start Reviewing.” You will be presented with your first flashcard. What I do is to try and read the sentence aloud as best I can and then click on “Show Answer” (or press the spacebar). Then I check and see if what I said matches up with the reading, in addition to how well I understood the contents of the sentence. You have four options based on how well you remember – Again, Hard, Good, Easy. If I made a mistake saying or understanding any part of the sentence, I click Again. If I struggled but got it right, I click Hard. If I went through the sentence without trouble, Good. If I fly through the sentence, speaking and understanding nigh-fluently, I click Easy. I will usually try and study my deck twice a day in order to better take advantage of the nature of the spacing of the flashcards.
One last thing as far as Anki itself goes, which is synchronization. Anki has a fully-functioning online version that allows you to study in a web browser by creating an account and syncing the deck on your computer to the Anki server. There are also some other interesting applications of this feature. Let’s get that set up now. First, go to Settings -> Preferences:
A window will pop up, under which you should see a tab called “Network.” Select it.
Click on “Create a free account” to be taken to the Anki website where you can set up your new Anki account. Once you have it created, type your user name and password into their respective fields in this window and make sure “Sync on open” and “Sync on close” are selected. This makes sure that you always have the latest, most up-to-date set of flashcards in your deck and on the server. After you do all this, click “Close” and your deck will now be able to sync! In the event of catastrophic computer failure, you can also sync with the Anki server to recover your deck, which will save you a lot of heartache in the long run.
Now for a few tips on sentence mining. The most important thing to make these flashcards enjoyable to read it to keep it short. Japanese sentences have the tendency to be many lines long, and if there is multiple new words in it, then you will almost certainly get frustrated remembering everything. I try to keep it to no more than two, maybe three unknown words in a sentence, or two lines long in the “Expression” field, which ever comes first. In this way, I’m able to quickly read the sentences in my flashcard deck and make more efficient use of my time. Of course, you may have a sentence that you want to mine that is really long, such as in a news article. At this point, it gets a bit more fuzzy and requires you to have a bit of a handle on Japanese, but the best thing I find it to cut up a sentence at “、” symbols, which are Japanese commas. If the sentence contains none of these, stopping at certain particles (は、が、に、etc.) allows you to break up a sentence while keeping it intelligible.
I know that I mentioned that you should stick to copy-pastable items, but if you have a manga or movie that you are dying to watch/read, then by all means go ahead. You may still find the internet to be of use to you. Once you come across an unknown sentence, Google it. More often than not, somebody out there has transcribed the book/movie in question. Simply copy-paste your sentences from there instead of manually typing them in. The most important thing is to not get bogged down, which you are like to do if you are just starting out. Keep it simple, short, fast, copy-pastable.
Finally, if you have a iPhone/iPod touch and are willing to jailbreak, check out AnkiMini. It’s a mobile version of Anki that let’s you study your deck on the go by syncing up with your account on the Anki server (aren’t you glad you set up an Anki account?). This is currently the best solution on the iPhone/iPod touch until the developers create a proper App Store app.
Well, that’s that. I hope that this guide will give you head start on sentence mining. If you have any further questions, comments, etc., please don’t hesitate to leave a comment!
Themocracy
That’s an absolutely phenomenal explanation.
I’ve been using smart.fm, but may have to give Anki another go.
It’s a very good explanation, but where do you start if you’re just breaking out of Heisig and you don’t understand *any* of it? I can read a few Japanese words here and there from habit (お休み, 日本語, 英語, etc.) but does it all flow like that? I don’t want to depend on Rikaichan the rest of my life. :)
@Rick Martin: Thanks for the compliment! I hope you get a chance to try Anki again. :)
@2nihon: I remember being back at the very beginning, it was quite grueling. :/ You have to find simple materials to read and listen to. I’d suggest grabbing the Minna No Nihongo textbooks, they are in pure Japanese and start you off with the simplest of sentences. Even though it might not be as fun as reading manga or something, it’s definitely a lot more fun than being stuck on a single page for an hour and still not understand anything. For grammar, check out Tae Kim’s grammar guide at http://www.guidetojapanese.org/.
As for listening, I find that “Learn English” podcasts for Japanese people are quite easy to listen to, since they are talking about basic English. You will pick up a ton of neat (but essential) vocabulary and phrases this way. I’d suggest http://www.englishpod101.com to this end. You can look it up on iTunes. Just keep at it! Stuff will start to flow, you just have to give it time. :)
I’ve just started using Anki recently for Japanese, and your article is a big help. I’ll also give the sentence mining a shot. :)
[...] available and the best on of them for me is the Japanese one. Visit Joe In Japan’s blog post Sentence Mining Explained for a good tutorial on how to install and use [...]
There’s also an iAnki plugin for iPod/iPhone. You can install it from Shared Plugins through Anki. It doesn’t require a jailbreak, but it’s not as good as AnkiMini.
Interesting! I know there are people out there that want to use Anki on their iPhones and iTouches but don’t want to jailbreak, this should come in handy. Thanks for the tip!
Awesome explanation!
I’m following the ajatt method as well but I’ve always been intimidated by sentence mining either for the fact that I didn’t understand the steps behind it or that I wanted to build my vocabulary first. Your explanation took that away.
Also, the tip for mining from Minna no nihongo is a great one, so I wanted to off source for the people who don’t have the funds to buy the book.
http://www.studyjapanese.org has an archive of terms and sentences from many Japanese textbooks. Minna no nihongo is one of them. Just go to the site, then on the left side bar click on Flash Card Trainer and look for the book in question. There is Minna no nihongo 1 and 2 as well as many others. The sentences for M no N seem to be listed at the bottom of the lists.
Again, thanks a bunch Joe!
[...] available and the best on of them for me is the Japanese one. Visit Joe In Japan’s blog post Sentence Mining Explained for a good tutorial on how to install and use [...]