[ファブリス]

For those experimenting with "kanji chains", there is an index on onyomi groups that I was working on a while ago.

See the news archive On (Chinese) readings of kanji.

You can also try KanjiTown, a method developed based on that paper I believe. In RTK2 there is a long phonetically sorted list of on-yomi kanjis. With this list you can create stories in a fictional setting, KanjiTown. you make each story based on kanjis with the same sound, and you make sure to place these kanjis in the same location in your fictional town. I placed AKU (includes Bad and Grip) at the KAnjiTown Akumulator (the powerplant) and made a story about how it went BAD (boom) and firefighters are hanging by a GRIP with their fingers onto the roof of the building.

I believe RTK could carry a RTK2 section, where you list the chains of similar sounding on-yomi kanjis, and people share stories on chains to remember them by.

[ファブリス]

Immacolata Wrote:I believe RTK could carry a RTK2 section, where you list the chains of similar sounding on-yomi kanjis, and people share stories on chains to remember them by.

Well that's the list. It was built with only the rtj1+3 kanji, and completed with the kanjidic info. I used the first ON from the kanjidic as the main one. Later Member 'kurojohn' who completed rtk1+3 went over it and changed some of the ON readings with the ones from RTK2. There are a few differences with RTK2 but not many.

But I'm not sure about sharing stories for the kanji chains. That would be highly variable depending on what stories people chose for the individual kanji.

What I had in mind is allow people to share 'tags' and short description for the actual ON groups. There are about 200 groups so it's worthwhile. Then you could add groups one by one, as you learn them, and review readings with all the 'known' groups shuffled.

Ah yes, I see now what you intend. I'm building RTK2 kanjitown stories now, and naturally they stick to RTK1 ones. I would have to go back and redo them all when I finish RTK3 some day.

Spotted an omission error btw. The group アン　does not include 行. Often known as the いく　kanji, it also has an onyomi アン.

[ファブリス]

The kanji you mention is probably placed in another ON group. The list only includes each kanji once. If it's to be used for kanji chains, the idea is to keep things simple and add the additional readings as you read Japanese material.

I've only made 4 chains so far for about 100 readings. The longer ones are actually easier. I haven't experienced the very small groups yet. Groups of 1 kanji per reading become equivalent to rote memorisation..

One idea was to create a special location for all the 'orphan' kanji (1 kanji per reading), couples (2 kanji per reading), 3 & 4 maybe. To help with the small groups.

[ファブリス]

Immacolata Wrote:I would have to go back and redo them all when I finish RTK3 some day.

When you make those chains you can always add more to them later. You don't have to restart from scratch. You can spend some time 'cross linking' back and forth between the existing kanji in your group and the ones you add.

行 is listed in the group コウ

Does it really matter? I mean, the method for using multiple on-yomi is to just include it in different locations for each yomi. So you should make a note of it and put 行 into your あん chain as well.

I have worked now with kanjis that have several onyomis. Works just fine. You do not confuse them, but when you run across a word where the kanjis have several onyomis possible, well you rake your brain for as many yomis as you can think of. You will soon get a feeling for which ones are common and which ones are rarer.

I think it is a mistake that you only include them one time. If you want to do the kanjitown method, you need the whole list. There must be an entry for every kanji's onyomi. And it does not destroy your stories. I just ran across 依 and 衣 in both my イ　and my エ stories. No biggie there, mate. And really, what is the greater idea in trying to keep things simple when it IS in reality complicated. Then it reveals a problem with the kanjichain method, because it works fine in KanjiTown... There ARE many kanjis with multiple readings in japanese, so deal with it rather than try to cut away readings that you ultimately force people to take up later as stragglers that got lost in the proces.

I have an excel file which is based on RTK2, it is made by leo smith, if you are interested.

[ファブリス]

I would have to rework my Perl scripts and see just how many more kanji you get overall if they are duplicated for every possible ON reading. If I limit myself to one reading per kanji then it's like RTK I, there will be no more than 2042 kanji to "chain". I'm not against duplicating the kanji though. I will have to udpate my onyomi list.

[ファブリス]

Ahh, 行 is a good kanji for this discussion.

It does have the reading アン, and its exemplary compound in RTK II is "a religious pilgrimage" 【あんぎゃ】

However in the Kanjidic database, and in Kodansha Kanji Learners Dictionary, the readings are listed in this order : コウ, ギョウ, アン.

The first two readings are both common, and I agree that in this case it is useful to learn both readings. However the アン reading is much less used, and could be learned from compounds, or be learned separately only when needed.

----

Another example,this time kanjidic is not helpful :

窪 "depression" heisig: 1323

In RTK II it is listed in the group "characters with no chinese readings".

The character does not appear in KKLD dictionary.

It IS listed with the readings ワ and ア in the kanjidic database.

The dictionary search gives 322 matches, 318 of which are names, nearly all using the kunyomi reading, and the 4 remaining dont use the chinese readings either :

窪み 【くぼみ】 (n) hollow, cavity, dent, (P)
盆の窪 【ぼんのくぼ】 (n) hollow at nape of the neck
窪む 【くぼむ】 (v5m) to cave in, to sink
窪地 【くぼち】 (n) pit, hollow, basin, depression

So I guess I'll have to filter out the readings no listed in RTK II or KKLD.

It seems that clicking on many of these groups (especially on most of those with just 1 kanji listed) results in an "invalid data" error in Firefox... Is it a bug?

laxxy Wrote:It seems that clicking on many of these groups (especially on most of those with just 1 kanji listed) results in an "invalid data" error in Firefox... Is it a bug?

Have a look at this thread.

Anyway laxxy, I thought you were a Twinkle man!? (In that regard, I've posted on the Twinkle site about problems I'm having, any chance you could help me out?)

Cheers,

synewave Wrote:

laxxy Wrote:It seems that clicking on many of these groups (especially on most of those with just 1 kanji listed) results in an "invalid data" error in Firefox... Is it a bug?

Have a look at this thread.

What I see there does not look like the problem I am having though...

Quote:Anyway laxxy, I thought you were a Twinkle man!? (In that regard, I've posted on the Twinkle site about problems I'm having, any chance you could help me out?)

Cheers,

I am -- but I got curious about how reading-based kanji grouping would work for me, so I was browsing around that part. It might be more effective than I used to think.

I'll take a look at the Twinkle site -- where are your questions there? forums, or one of the trackers? sf.net interface is rather horrid.

I see your message, I'll reply there too, but here is some advice, perhaps some of it could be of use --

1. You will not see Japanese characters under an English OS Palm anywhere except the card review screen itself. You will need to use TwinkEdit to edit the cards (or install J-OS or the like). This is exactly why I was suggesting that getting a used Japanese Palm could be a good idea. It does not bother me much though (and I actually have edited cards on several occasions, you just need to use your intuition

2. It is essential to set *both* the font to Japanese and the encoding to EUC (or whatever) in Twinkle.

3. You should not mix Japanese and English characters in the Japanese fields. Note that the Japanese IME will often enter English characters -- such as "English" numbers and symbols "123~,..." etc. instead of "１２３～、。?" etc. You should not let it do that -- they will display but they are not going to look nice. Generally it is best to avoid fancy characters, but you can try and test which work and which don't.

4. It is strange to see that your periods etc. appear high -- could it be a font issue? Mine display properly (I'm using a low-res font), and I haven't noticed any irregularities when I (briefly) tried it with a high-res font on my wife's Tungsten E2. There are several high-res fonts there, maybe you can try some other one?

Have you resolved the issue with the characters being small and of bad quality? With high-res fonts they should look pretty good (and quite decent with medium-res too).
BTW, do install PADict, it uses the same fonts and is quite an essential program. You'll also be able to check if you got everything right with the fonts that way too.

synewave Wrote:

laxxy Wrote:It seems that clicking on many of these groups (especially on most of those with just 1 kanji listed) results in an "invalid data" error in Firefox... Is it a bug?

Have a look at this thread.

I am having the same problem with "invalid data" pages in Opera, and it isn't the same problem as the one in synewave's screenshot (actually, the readings with errors in his screenshot aren't even listed on the onyomi page for me to click in the first place).
These invalid pages occur quite frequently. For example, the first row is 50% error pages, more specifically when clicking these readings:
アア アオ イエ イジ イウ
This is unfortunate as the onyomi groups with few characters are the only ones I want to study using this mnemonic method.

I tried viewing this. Anything I click gets "mydb::query failed" at the top of the screen and I don't see any kanji in any of the groups. Am I doing something wrong?

same here. I think maybe it's down for the moment.

hm...and now it looks like the page doesn't exist. Fabrice, any chance of re-upping this, even with a "This is pure super beta alpha" notice? It would help just to have that list available in a convenient spot! Thanks!

another request to put this thing back up - I'm really needing a centralized list of all of the RTK 1 + 3 kanji, sorted by 音読み!

(or if someone has something like that in some sort of normal document format, that also would be fine!)

You know, that would have been really useful for me a couple months ago...

Yeah. I actually posted this message just after (finally) downloading the RTK1 + RTK2 Numbers.rtf you referred to earlier. If everything was ordered in Onyomi, it'd be fine.

As an aside - if anyone is just looking for the RTK1 Kanji ordered by Onyomi, you can check out this post:

http://my.opera.com/Immacolata/blog/show.dml/316801

It's got the whole KanjiChain/Town/Whatever thing going on, so the story column I'm not much interested in, but it may be useful for someone.

[ファブリス]

The closest to that that I have is a list compiled by John Vold.

Code:

# Index of 3007 Heisig-numbered kanji with On-yomi

ア    阿    1295    Africa
    亜    1809    Asia
    蛙    2693    frog
    唖    2958    babble

アイ    哀    401    pathetic
    愛    737    love
    娃    2200    fair
    挨    2248    shove
    曖    2437    equivocal

アク    握    1059    grip
    悪    1810    bad
    渥    2320    moisten

アツ    圧    152    pressure
    斡    2781    auspices

アン    安    190    relax
    案    214    plan
    暗    480    darkness

(etc...)

He hand edited it so only the first line of each group contains the OnYomi. So if you want to resort on the OnYomi for example, you'd have to duplicate the value in the empty cells. It's UTF8 and tab separated. I have two versions, looks like one is John Vold's selection, where he manually picked the appropriate group for kanji that has several readings; another version seems to be automated from KANJIDIC. You'll have to figure out the differences.

Email me from the forum or contact page on the main site and I'll send you a zip file with some data files I have.

ファブリス先生、ありがとう。

Been away for a second, just now getting back online. Check your inbox.

Hey Fabrice, would you mind sending me that Onyomi reading index file.
I started making my own yesterday, based on the RTK2 index, but it was taking waaaayyy longer than I expected it to. I spent about 4-5 solid hours working on it and only got about 15% done. I dont really wanna spend another 20hrs working on it, so if you still have it lying around, I would Love to get that from you.

my email is maikerusan at gmail dot com

thanks
michael