Chart of Frequency of Kana used in WK Kanji readings

I read vargsvans’ great post shortly after I started my WK journey, and I found as I progressed that like them I was having increasing trouble with some of the kanji mnemonics, especially in distinguishing between long and short vowels, so I took vargsvans’ suggestion and read Moonwalking with Einstein.

Armed with the information I learned from that book, I am now in the process of reworking the mnemonics to better suit the way my mind works (I really like consistancy, I really don’t do well with gross humour/violence), with a focus on the onyomi [edit: 1 & 2 mora readings] since most of the main readings for the kanji are onyomi, and to that end I asked rfindley to crunch the data for me, which they kindly did with alacrity!

So, in case anyone else is interested in how often they will need to remember particular kana/mora/syllables/sounds, I’ve made a chart. I was interested to find that the combos (with little や、ゆ、よ)only occur singly / in first position, never in second…

For simplicity, if a mora didn’t show up at all, I removed it (the whole ぱ line is gone) and I didn’t bother entering any zeros.

28 Likes

Very interesting! Thank you for sharing!

2 Likes

Ooo, love that data. Kind of interesting the sorts of things that ever appear as the second mora.

Hear hear. The number of mnemonics that involve toilet humour or violence just one inch this side of kicking puppies. Helps me to remember the readings sure, but then I feel too disgusted to actually type them. Like the mnemonic for 質…

2 Likes

Many thanks to you and @rfindley for this data! I’m a numbers kind of guy, and this type of thing is super interesting to me. I’ve been wanting to design a kanji learning system based more on the phonetic system, so this is quite useful to know! :grinning:

Question: What has a first mora reading of づ? The only examples that come to mind are all suffixes (cases of rendaku).

2 Likes

Not entirely sure if “with a focus on the onyomi” means the data includes only on’yomi, but has づ listed as an alternate kun’yomi.

1 Like

The data includes Wanikani’s “accepted answers” for each kanji. So, some are on’yomi, some are kun’yomi.
Also, it excludes kanji with 3 or more mora.

1 Like

~づけ - pickled, lvl 59, apparently!

EDIT: That’s a vocab item , not kanji, so I’m not too sure now…

I suppose that answers my question with 詰 then, as its kanji requirements are the 訓読み, つ and づ, which slightly bugs me now, as my dictionary does not list づ as a reading outside of rendaku as a suffix. Makes me wonder why 付 and 漬 don’t allow づ then. :thinking:

EDIT: Spoke too soon. So it was 漬 instead. Regardless, it’s the same policy that slightly irks me. Oh well! :laughing:

DOUBLE EDIT: If it is the accepted answers, then it must be 詰. WK’s accepted answers for it are つ and づ. While 漬 does have the suffix づけ as vocabulary, its sole acceptable reading is つ, making it impossible based on what is stated above. Either way, I’m sad by WK’s inconsistency. :pensive:

Neither am I, frankly - I asked for rfindley for data on readings with only one or two mora with the understanding (??) that 3+ mora would be kunyomi, but I expect that there may be some shorter kunyomi in the mix. Perhaps I could have specifically requested onyomi… I thought I was asking for something quite complicated, but was stunned with how quickly rfindley wrote the script and got the results (less than 2 hrs after my request)!

It’s actually quite easy to process such things now with the new APIv2 and Wanikani Open Framework. The hardest part (which wasn’t hard) was writing the line of code that extracted the mora.

If you want, I can run it again with only on’yomi when I’m at my computer later. It will take less than a minute since I’d just have to add a filter for 'yomi type.

1 Like

I think what you have done already is fine for me personally, excepting I may have ‘misadvertised’ it in the top post here. I wanted to get a better picture of the yomi across WK that weren’t the everyday words I already knew (there were only 6 3-mora words in the first 10 levels of kanji, and I already knew them all, unlike many of the 1- and 2-mora yomi). :blush:

I wanted to see it sorted by frequency, so I took all the values from your chart and put them into a spreadsheet. If anyone else is interested in it that way, here it is in comma-separated format:

Single/First Mora

Single/First Mora,Frequency
か,146
こ,103
し,83
せ,78
は,76
しょ,73
さ,70
そ,60
け,58
と,55
た,52
き,49
い,47
ふ,43
て,40
しゅ,38
ぼ,37
ほ,36
きょ,36
え,35
ひ,35
よ,34
す,32
お,30
あ,29
じょ,29
ゆ,28
じゅ,28
が,27
じ,27
ちょ,27
れ,26
だ,25
ご,24
も,24
りょ,24
く,23
ば,23
げ,22
ま,22
ち,21
ど,20
な,20
つ,19
り,19
ろ,19
う,17
きゅ,17
ぞ,16
へ,16
や,16
しゃ,16
み,15
め,15
わ,15
ら,14
ぎ,13
ぶ,13
に,12
ね,12
ざ,11
の,11
ちゅ,10
りゅ,10
ぜ,9
む,9
ひょ,9
び,8
ぐ,7
る,7
ぎょ,7
べ,6
びょ,5
で,4
みょ,4
ず,3
きゃ,3
じゃ,3
にゅ,3
ぬ,2
ぎゃ,2
ちゃ,2
づ,1
ぎゅ,1
にょ,1
ひゃ,1
みゃ,1
りゃ,1

Second Mora

Second Mora,Frequency
う,491
ん,387
い,245
く,171
つ,104
き,46
ま,15
た,14
か,10
ち,10
な,10
ら,10
ろ,10
さ,8
り,8
る,8
わ,8
し,7
め,7
ば,6
み,6
こ,5
と,5
ど,5
に,5
も,5
け,4
が,4
ぎ,4
じ,4
ね,4
れ,4
お,3
す,3
の,3
び,3
や,3
え,2
せ,2
そ,2
て,2
だ,2
べ,2
ず,1
ぞ,1
で,1
ぬ,1
は,1
ふ,1
ぶ,1
ぼ,1
む,1
ゆ,1
よ,1

…or you could have followed this link in the original post!

I’m in awe of everybody who can manipulate data like this - perhaps I’ll bother you instead of rfindley next time! :blush:

doh! :woman_facepalming:
I guess I will leave it here in case there are any other lazy people who can’t be bothered to follow links…

Actually, all I did was data entry from your pretty chart; I have no idea how rfindley got the original numbers!

1 Like

The java script is there too, and I actually dared to look - while I could never write it myself, knowing what it did, it seemed a much more ‘natural’ language and so more understandable than the wee bit of ‘programming’ I did in the late 80s.

I just remembered I got pinged by this thread at a time where I couldn’t check it out properly.

Just wanted to say your system looks really impressive! I’m happy to hear you got something out of reading the book. Knowing the frequency of a reading also makes it so you can focus your most memorable mnemonic devices to the readings where they have the biggest impact, that’s really cool :star_struck:

1 Like

I was initially stumped as to how to apply the memory techniques from the book until I realised that if I focussed on one and two mora yomi (which appear to be the vast number of kanji readings), I could have each single/first mora as a person (family, friends, famous people/characters) and any second mora, of which there are far fewer, as an object strongly associated with that person. So, for example, ら is: 1st Raj Koothrappali, 2nd his Yorkie Cinnamon, and く is: 1st Cookie Monster, 2nd a cookie - this then makes the kanji yomi らくto be Raj with a cookie, something more memorable to me than a generic ‘rack’.

I’m just about finished assigning mnemonics (some of the combos are proving tricky) but I hope to be resetting soon and getting back to daily effective study.

Thanks again for taking the time to write that wonderful post :blush:

3 Likes

I’m with you on not liking the violent mnemonics. I’m about to go search for a user-script that can actually change the mnemonic already given instead of just adding notes. :+1:

Please let us know here if you find one!

I do remember that rfindley wrote some stuff to allow folks to highlight their notes (blue for rads, pink for kanji, purple for vocab) to make them look as good as the WK mnemonic notes - let me see if I can dig that up…

Found it!

2 Likes

Thank you for finding that!

Sadly, I couldn’t find the mnemonic changing user script. But I’ll keep an eye out!

1 Like