Learn Japanese via your favourite Light Novel Bookclub?

[baerrach]

I've fallen in love with reading Accel World and Sword Art Online, and since they now have official English print versions are abandoned projects.
However the official publication rate is too slow and personally I would rather my money went directly to they author than in an around-a-bout way.
I also (crazily) figure I can learn Japanese and read these in the original form about as fast as they would be published in English, with the added benefit of being able to read the newest books as soon as they come out.

So I am currently following the insane path of learning Japanese so I can read the original books (I have the first three Accel World on my desk).

I'm currently working my way through the Core 2000 Anki decks but that isn't yet getting me closer to being able to read the books.
I'm about to start looking into building an Anki deck based on word frequencies and page order so I can practice on my desired target.

I've found one or two other people attempting similar things for other books on the net, but they already know Japanese and are trying to use their favourite book to help others learn Japanese, and their efforts dry up pretty quickly. I'm realistic enough to know I'm most likely to give up at some stage because of the amount of effort involved, but my OCD won't let me stop right now, and while its still driving me along it might help someone else's effort to learn to read Japanese. And I have more incentive to make it happen because I can't read the books myself unless I have the tools.

Do you think a bookclub like this would be suitable on baka-tsuki?
Could abandoned project pages be used to coordinate people wanting to learn in such a fashion?
It's possible this may produce some more translators for other projects in the future...

Cheers
Bae

[rydenius]

Hi Bae,

Not really sure I understand the book club idea, since it basically sounds like what Baka-Tsuki is all about anyway... but since I'm doing something similar myself (attempting translating light novels to improve my Japanese skills) I can offer some advice on getting your reading and writing skills up to the level you need to begin translating the novels. (It's a crazy amount of work and will require dedication, but it is achievable!)

Assuming you are starting from no knowledge of Japanese, the first thing you are going to need is an understanding of grammar (mainly via particles) and verb conjugations. With Kanji-based words it's relatively simple (just look them up in an online dictionary -- once you know the radicals), but for the hiragana that makes up the verb endings (conjugations) and other grammatical particles, it's impossible to translate without knowing enough of the grammar to be able to figure out what you even need to look up. Also most dictionaries will only have the verbs in their dictionary form and online translators are almost completely useless (except for the occasional "set phrase"), so if you don't even know the verb conjugations, you are really stuck.

This is a really long post so I'm going to put my recommendations inside a spoiler tag:

Spoiler! :

For getting started on the grammar side of things, I recommend the Genki I and Genki II textbooks. If you just focus on the Grammar and story sections you can work through the material fairly quickly (probably plausible to complete the grammar sections of both books within 2 months time). Then I would follow that up with Essential Japanese Grammar by Masahiro Tanimori and Eriko Sato (this is also a great reference book that I still get tons of use out of). Which will get you up to the Intermediate level were you can progress to more difficult textbooks such as: An integrated approach to Intermediate Japanese. If you're still struggling to get to the Intermediate level, I also recommend Fundamentals of Japanese Grammar by Yuki Johnson, though it's a bit drier read than Essentials of Japanese Grammar. I tried a lot of books when I got stuck at the upper beginner level, and these last two books were what finally allowed me to bridge the gap from beginner to intermediate.

If you can't afford the textbooks, a good free resource to start with is Tae Kim's Guide to Learning Japanese: http://www.guidetojapanese.org/learn/grammar
And Michiel Kamermans's e-book (also available from Amazon) has in-depth explanations of verb conjugations that really gets into their underlying workings:
http://pomax.github.io/nrGrammar/


As far as vocabulary goes, the Anki Core 2000 decks seem like they'd be much too hard to start with. But it's great that you're already doing SRS with a flash card app. To that end, you'll want to check out memrise -- it's a free community-based learning site with lots of helpful flash card style courses.

If you don't have all the hiragana and katakana learned, you'll want to do that first. Using flashcard decks w/ audio files (and practicing writing as well), it shouldn't take more than a few days -- one week tops. This step is vitally important! (If you're feeling ambitious you can lean some basic vocab to help with remembering some of the kana, but it's not really needed.) Also don't forget to set up your IME and get used to typing in Japanese kana during this process.

Next learn the English meanings for ALL 214 KangXi radicals (and how to write them): http://en.wikipedia.org/wiki/Kangxi_radical There are good Anki and Memrise decks for this that'll get it done with no problems. Don't worry about the Japanese names for the radicals unless you want to challenge yourself. The importance here is that you'll be able to build stories and memes based on the radicals that will make memorizing (and retaining) a particular kanji's meaning much easier. Also it's important in being able to look up a particular kanji by using its radicals.

Next you should start on learning meanings of the grade 1-3 Kanji (as much as you can without feeling overwhelmed) -- but don't worry about the readings. The readings are useless without knowing the vocabulary word(s) that they go with. The trick here is that you should also begin learning the JLPT N5 vocabulary (or the Genki vocab lists -- though some of the early words in the Genki list are quite difficult for a beginner), and when you learn the Kanji associated with the vocabulary word and it's reading, you'll better be able to remember the readings because they'll be associated with an actual vocabulary word, rather than just an abstract Kanji.

As you learn new Kanji and vocabulary, make sure to look them up in both:
http://beta.jisho.org/ (the public beta version of the Jisho.org site -- it's much improved over the old version!)
http://www.alc.co.jp/ (this dictionary has tons of example sentences so you can see the vocab in context. It also has no problems with partial words or grammar, and is great for looking up phrases. Just type in the search field, then press enter or click the yellow button.)
This way you'll be able get a better grasp on the context, usage, and collocations.

From there you can go on to learn the rest of the 2136 jōyō kanji and proceed through the vocab for the other N levels (N4, N3, etc.). Personally I'm using S4Nii's "2136 Jōyō Kanji by Grade" course on memrise to do this myself, and I've gone from knowing the meanings for about 500 Kanji to knowing the meanings for over 1280 Kanji in a period of less than a year (from my own experience, it's certainly feasible to learn all 2136 in less than a year). Another advantage of S4Nii's course is that it's based on the Jisho.org definitions, so (even if you don't know the readings) you can often quickly find the kanji you need to look up a particular word just by searching for the definition keyword. For N5 and N4 vocab, I'm using the "JLPT N5 Vocab" and "JLPT N4 Vocab" memrise courses by jlptbootcamp. I finished N5 last year, and will be finishing N4 in the next week as I continue to prepare for the N4 exam in December. Not only is there sample audio (some of dubious quality though... ) with each vocab word for the N5/N4 courses, often you can find useful information in the "mems" that other users have created for the various vocab entries.

When you need to take a break from cramming vocab, kanji, and grammar, watching anime or Japanese dramas can help with both de-stressing and reinforcing what you've learned. Having audio and visual memories to draw upon is also helpful when creating memorable "mems" for remembering vocab or kanji.

From my experience so far, the N5 level isn't used much in light novels, but once I got a ways through the N4 level, I started to find that I didn't have to look up words as much. This is mainly a speed convenience because it can be _really_ slow going if you have to look up every word... and that can be very discouraging. When I first started out, recognizing what was a grammar particle and what was verb conjugation (and what was a word or mimetics in kana form) was the hardest stumbling block, but it does get easier with practice. Considering how slow going translating is at my level, I translate by breaking the sentences down into their grammatical phrases and typing them directly into a text editor (MS OneNote in my case), so that I can revise the English as I go and I can store it for reading, or further revision should a later section clarify problems in a previous section. Considering how much help reaching N4 has been, I imagine that the achieving the N3 level will be an order of magnitude even more significant, but that's going to be a ways off for me yet.

I hope that's helpful to you.

Personally I still consider myself a newbie translator and the other translators on this site are going to be much more skilled than I am, but if you get stuck and need help figuring out a particular grammar construct, I certainly don't mind helping. And maybe some of the others would help as well. Posting snippets here that you're having trouble with should be okay, but since it's licensed, keep in mind that continuing the project on Baka-Tsuki is not going to be approved. There's nothing stopping you from posting your translations on your own website or blog, though -- but you'll want to be careful not to link to it from Baka-Tsuki.

As far as using the forum to create a resource for beginning translators, I remember early on that there were quite a few constructs that I couldn't find information about no matter how hard I looked, but later figured out as not so obvious variations of usages covered in my textbooks (or contractions of usages contained in the textbook ). The advantage to being a novice is that you can still identify those constructs that are "hard", so it could certainly be useful to create a listing of problematic grammar in order to use as a reference and to help others from getting stuck. I think it would probably be okay to use this thread to do that and to get help with your translation if you wanted.

You could also use the Auxiliary Brigade Translation Check thread here:
viewtopic.php?f=44&t=6078

[baerrach]

I appreciate the reply, including its length.

I'll try to expand on my bookclub idea.
And I'll give a little background as well.

Background first:

Currently I'm a only just a little past the first step in learning Japanese.
I've read Tae Kim, some AJATT stuff, and a few others I can find online.
But I definitely don't have a grasp of grammar or verb conjugation.
It's all intuition at the moment on how the sentence feels.

I've already learning the hiragana and katakana, and can probably handle with ~80% correctness the Core 2000 Step 1 Anki deck (https://ankiweb.net/shared/info/4271439873)
That's the 1000 cards they have; being able to recall the japanese from english, read the japanese vocab, and read the japanese sentence vocab.

I've also just borrowed whatever the library had on japanese courses and I'll see how that goes.

Expanding on the idea:

So instead of just building my vocab from whatever happens to be in the Core 2000 decks I figured it is worth building my own deck, a deck that contains words from the light novel that I want to read.
But such a deck, of course, does not exist so I need to create it.

I've bought the books, but I'm lazy and my day jobs is programming so I've grabbed a copy of the raw images to save me doing the photocopying and I'm using tesseract (https://code.google.com/p/tesseract-ocr/) to give me the japanese text instead of an image. It's not great at about 80% accuracy but doing verification and validation plus the odd correction is certainly faster than typing the book by hand. Especially since I dont know the radicals yet and can't look up the kanji anyway! With a little bit more investigation I think I can improve the accuracy rate as well.

I can now start building my deck from the OCR'ed text. I haven't given a lot of thought to what the correct order of the deck would be. I'm currently leaning to the lexical order of the book.
i.e. start at sentence one, word one, and build cards for each word. Each new word would get its own card and be entered only on the first occurrence. As I progress there should be less words needed for cards.
This way the vocab I am learning at least lets me read the book and in page order. Other options would include word frequencies, or sorted by jōyō order.
Of course I'd be wanting to automate that process, but I'll need to understand grammar better in order to be able to work out word boundaries...

The translations I would be doing would be literal rather than english'ifying the whole sentence, and it wouldn't be every sentence in the whole book, so I don't think it should fall under the abandoned rule.
It should fall in the grey area the site mentions in the FAQ.
I'm sure there are a small number of like minded insane people who would find an Anki deck on their favourite light novel appealing.
Copyright varies from country to country, and I'm not a lawyer, but I figure this falls under research.

And if I can manage to build such a deck then I'd like to be able to share it (that's the grey area legal bit).
If people own a copy of the book then it should be less of a legal issue as it's personal use.
If I can work out how, then I'd like to let other people collaborate as well.

If it is suitable, then I think the best place for it is on baka-tsuki as this site has the best exposure for finding other like minded people.
This might also be a way to cultivate more translators.

If the process is reasonably painless, then it could also be used on other light novels.

Who would be able to give definitive answers on the suitability of such an idea?

[rydenius]

Ah I see what you mean. You want to make a vocab/sentence deck from the novel. I actually started doing something similar with the light novels that I've been working on. I wanted to create a more useful "commonly used" words deck, since at the time the existing decks didn't seem that useful to reading real Japanese. Eventually I stopped working on the database because it was taking away from my translation time. If you could use OCR and parse out the vocab, that would greatly automate the process but as you point out here:

but I'll need to understand grammar better in order to be able to work out word boundaries...

And this is the rub. In order to work out the word boundaries, you basically have to do the translation.
I do have some ideas for writing a small application that would allow a user step through a text, manually select the word boundaries (or the words themselves), and add the words into the database. Incrementing the occurrence count of each (duplicate) word as it's entered into the database would also help with evaluating and sorting vocab items by how often they are used in real contemporary written Japanese works.

About AJATT, I think it's a theory with some benefits and it was also the direction that I initially attempted, but it reminds me of the "whole language" vs. "phonics" debate with regards to learning English and seems to have similar drawbacks. For me, I find that a functional grammar approach that focuses on how the various particles build the functional parts of speech (Adverbials, adjectivals, nouns, direct object, subject, etc.), is needed to build up the foundation from which to get the most out of the AJATT/"whole language"-style studying. Understanding how the meaning of a phrase is derived from its component parts would surely be more useful in the long run than memorizing English equivalents of Japanese phrases -- which in theory would better lead to the end goal of AJATT anyway: Using Japanese to learn Japanese.

As far as finding out suitablilty, you'd need to get in touch with one of these guys (from the Admin section of the front page):
thelastguardian
Onizuka-GTO
Vaelis (FR)
Bicube
Darkoneko (FR)
Teh_Ping(GPS)
KuroiHikari (GPS)

[capitas]

The advantage to being a novice is that you can still identify those constructs that are "hard", so it could certainly be useful to create a listing of problematic grammar in order to use as a reference and to help others from getting stuck. I think it would probably be okay to use this thread to do that and to get help with your translation if you wanted.