When I first started studying sentences I began to learn the readings for the Kanji. However, I noticed that when I was presented with a word formed out of compound Kanji (i.e. 高速道路) I wasn't memorizing the readings for each individual Kanji but rather the compound as a whole. When a Japanese is presented with a compound he/she have not encountered normally the first thing they try is guess at the readings with the Kanji.

After noticing this I began memorizing each reading for each Kanji regardless of its position in a word. Some times it helps at other times its just time consuming. Because when your presented with a compound word they don't break it up for you. Instead I have to go online and find a way to break it up... Is there a better solution to this?

I'd suggest just doing what you were doing in the first place and learning readings for words, not kanji. Eventually when you've done enough of that you'll be able to make guesses about readings when you haven't seen a word before... Guessing readings for individual kanji is not in my opinion a sufficiently useful skill to be worth putting extra effort in to obtain, when you'll pick it up more or less eventually anyway.

(高速道路 is actually a compound of two words 高速 + 道路 so breaking it down that far is probably worth doing.)

I had the same problem as you, and it went away after I started using japanese keywords with RTK. At first I just guessed the reading of that lucky kanji with a reading I knew, which helped me remember the other part of the compound, eventually I got used to assign each part of the reading to each kanji and it's been much easier since.

how far are you into the keyword-kanji method? and how has it helped overall?

I think it's good to know how each individual kanji is pronounced. I actually stopped using the furigana generation feature in Anki because I was annoyed that it didn't give the furigana for each individual kanji. Now I use Tobberoth's clean kana generation hack and then manually separate words and kanji readings. (This also takes less screen space than using furigana.)

Here's an example of what my reading/recognition cards look like:

Front:
(Image/snapshot.)
だから、人間は安心して眠ることが出来た。

Back:
(Audio.)
だから、　にん・げん　は　あん・しん　して　ねむ・る　こと　が　で・き・た。
だから (conj) so; therefore
にんげん【人間】 human being; man; person
あんしん【安心】(スル) relief; peace of mind
ねむる【眠る】 (v5r) to sleep (not necessarily lying down)
ことができる【ことが出来る】 (v1) can (do); to be able to (do)

I don't find that this really takes a lot of time to do. You can easily check/find readings using Rikaichan or Yahoo!辞書.

Edit: typo.

WOW thanks for the info shirokuro. are you saying that this plugin adds furigana to each kanji in the answer field? that would be soo helpful.

Sometimes when you learn kanji readings in a compound you learn the word as a whole and might not recognize the reading for one of the kanji in a different word when that kanji appears in a different word. I wouldn't worry about that though because there are so many words by the time you've learned several with that kanji in it you will have definitely picked up how to read that kanji. Really it's easiest to go with the whole word because you want to be able to pronounce the word instantly when you see it rather than thinking about how all the kanji are pronounced everytime you read it.

mastermx Wrote:how far are you into the keyword-kanji method? and how has it helped overall?

I must've been at 700-800 cards (including many I already knew well, from wrightak's deck) when I noticed I was actually paying attention to each individual reading. Besides that, it's taught me many readings (particularly onyomi) and, of course, reviewing RTK.

Funnily enough, it hasn't taught me much new vocabulary (Edit:vocabulary I've encountered) which I thought would be the biggest advantage over english keywords.

Are you using Anki, mastermx? If you are, there's a shared plugin called "Japanese Support" that automatically generates furigana for you. You can see a screenshot and find more info about it here: http://ichi2.net/anki/wiki/JapaneseSupport

It's a really great plugin. It sometimes doesn't get the furigana correct, so you have to check it, but it's still a huge time saver. I had been spending a lot of time manually changing the furigana it generated, though, because it generates it on a per-word basis instead of per-kanji basis. So I'd always change, for example, 日本語[にほんご] to 日[に]本[ほん]語[ご], probably 'cause I'm anal-retentive. XD Anyways, I ended up just not using it at all and typing in the kana manually for a while, but then Tobberoth hacked/modded the Japanese Support plugin to make it generate just kana, so I'm using that now. He posted about it here: http://forum.koohii.com/showthread.php?tid=4629

I agree with pm215 that you don't really need to do this. A lot of the time it'll be clear which kanji has which reading, and you won't need to look it up unless you want to be absolutely sure. You'll also probably just pick it up from learning lots of words that use the same kanji. But, for example, for words like 今日（きょう）, I think it's good to know that the compound as a whole is read きょう, not that it's 今（きょ）日（う）. So I do it this way, so that when I learn new words, I know that I'm getting the reading of each kanji or compound correct.

This problem disappears with time. You get a feeling for what is an on'yomi and a kun'yomi and in the same way, you get the ability to split words up automatically. There will always be slightly hard words (砂漠, is it sa-baku or saba-ku?) but in 99% of the time, it's very obvious. I'm sure pretty much everyone here who has studied Japanese for some time can easily tell that sa-baku is the correct split, simply because saba doesn't sound like a correct on'yomi. My advice is, don't worry about it.

mezbup Wrote:Sometimes when you learn kanji readings in a compound you learn the word as a whole and might not recognize the reading for one of the kanji in a different word when that kanji appears in a different word. I wouldn't worry about that though because there are so many words by the time you've learned several with that kanji in it you will have definitely picked up how to read that kanji.

I definitely agree that readings will sort themselves out eventually on their own, just from learning lots of vocabulary, reading stuff with furigana, and so on. You can't even avoid this; just learning vocabulary that use the same kanji cements readings in your head without you actively trying to memorize them. However, I still think it's better to try to get the readings right the first time you learn a new compound, because I think it's harder to relearn something correctly than to just learn it properly from the beginning.

Quote:Really it's easiest to go with the whole word because you want to be able to pronounce the word instantly when you see it rather than thinking about how all the kanji are pronounced everytime you read it.

It really only takes two seconds longer to find out how each kanji is read (hover over each kanji using Rikaichan and hit enter to bring up KANJIDIC, or look it up in 大辞泉). And I don't think that learning the readings for each individual kanji in a compound makes you think about the readings separately when reading, or at least that hasn't been my experience. I still read and recognize words as a whole, but I also usually know how each individual kanji is read.

Tobberoth Wrote:saba doesn't sound like a correct on'yomi.

There are rules for what can and cannot be an 音読み. They're not terribly interesting, but one of them is that if you have a CVCV pattern, the second vowel has to be i or u. Thus, さば can't be 音.

You can pick up the rules as you go along; nothing wrong with that, but if you're curious, I think they're on Wikipedia.

shirokuro Wrote:I don't think that learning the readings for each individual kanji in a compound makes you think about the readings separately when reading, or at least that hasn't been my experience. I still read and recognize words as a whole, but I also usually know how each individual kanji is read.

Ever since I've started learning the individual readings, I have gotta admit ive found it easier to remember compounds and non compounds alike. It seems like it would be a waste of all the time i spent on RTK to just ignore the readings for the kanji and view the word without breaking it down.

My current method of learning the readings involves typing the word in Japanese and seeing how Microsoft assembles the Kanji for each reading. If it does not give me the correct kanji for the syllables i give it, then I view it as an individual compound such as 今日（kyou)

mastermx Wrote:When I first started studying sentences I began to learn the readings for the Kanji. However, I noticed that when I was presented with a word formed out of compound Kanji (i.e. 高速道路) I wasn't memorizing the readings for each individual Kanji but rather the compound as a whole. When a Japanese is presented with a compound he/she have not encountered normally the first thing they try is guess at the readings with the Kanji.

After noticing this I began memorizing each reading for each Kanji regardless of its position in a word. Some times it helps at other times its just time consuming. Because when your presented with a compound word they don't break it up for you. Instead I have to go online and find a way to break it up... Is there a better solution to this?

What you did in your first paragraph is exactly what I did to learn kanji. I just stared Heisig to fill in any gaps in my learning, so the vast majority of kanji I know is simply from learning the readings to a word, and then seeing that specific kanji or word again so many times that I finally understood the reading for it. I never made any effort to learn individual readings -- in fact, whenever I tried, I would just find it boring and annoying and not be able to remember anything I studied.

The thing is that, especially for words like 高速道路, those kanji are used ALL. THE. TIME so it's impossible to escape from seeing them over and over and over again. Just learning vocabulary in and of itself essentially forces you to learn the readings. Once you become familiar with different vocabulary using these kanji, you will then be able to guess at the pronunciation of new words.

At some point, to, you start to just figure out naturally what reading is appropriate for each kanji and what is not. For example, 綺麗（き-れい）, 魔法（ま-ほう）, 掃除（そう-じ）, 時点（じ-てん）, et cetera. The idea that it's きれ for 綺, まほ for 魔, うじ for 除 and じて for 時 all sound ridiculous once you actually get used to how kanji compounds work and sound.

mastermx Wrote:

shirokuro Wrote:I don't think that learning the readings for each individual kanji in a compound makes you think about the readings separately when reading, or at least that hasn't been my experience. I still read and recognize words as a whole, but I also usually know how each individual kanji is read.

Ever since I've started learning the individual readings, I have gotta admit ive found it easier to remember compounds and non compounds alike. It seems like it would be a waste of all the time i spent on RTK to just ignore the readings for the kanji and view the word without breaking it down.

That's great! ^^ I really find that it helps, too. For example, definitely the most complicated kanji compound I have in my deck is 脊髄小脳変性症. Here's the card I have for it:

Front:
(Snapshot.)
脊髄小脳変性症。

Back:
(Audio.)
せき・ずい・しょう・のう・へん・せい・しょう。
せきずいしょうのうへんせいしょう【脊髄小脳変性症】 spinocerebellar degeneration
せきずいしょうのう【脊髄小脳】 spinocerebellar
せきずい【脊髄】 spinal cord
しょうのう【小脳】 cerebellum
へんせいしょう【変性症】 degeneration
へんせい【変性】 degeneration; denaturation; unfolding
しょう【症】 (n-suf) illness

Especially since I skipped RTK, I really don't think I'd be able to memorize this word if I hadn't had it broken up by kanji (and then broken up further into separate words/morphemes). The only kanji in it I'd known before were 小 and 変. I find it's a lot harder for me to associate kanji with words, or even to be able to read, if I don't know how the individual kanji are read. Also, readings sort of become hooks for me, which I find really helpful since I can't rely on English keywords, and then meanings kind of sort themselves out over time, just from learning more vocabulary.