[Userscript] Keisei 形声 Semantic-Phonetic Composition

@acm2010 I believe 省 is showing the wrong phonetic component. It says it’s associated with 生, but shouldn’t it be 少?

Seems like it previously looked like 眚, and it also has the reading せい, for example in 反省 「はんせい」. 省とは (ショウとは) [単語記事] - ニコニコ大百科

少 only has reading しょう available. I think you are right that they were thinking of that when the simplified the kanji, but some compounds still remain with the old readings. No harm in remembering 少 as a tone mark, though.

Could you do something to indicate when the tone mark is no longer in the present-day version of the kanji? I use this script to help me notice phonetic components that I may have overlooked. But it’s not realistic to know historical versions of kanji in order to correlate them to a phonetic component that is no longer visible in the kanji.

I can add another category to the badges to make it faster to see and more clear that a kanji is intentionally not fitting in. I will mark them when I come across them, so they will appear over time. There are already “non-compounds”, I added 省 to 少.

Sometimes the tone marks take more effort than being helpful, this kanji business is quite messy.


@seanblue @gzovak New version 1.1.5 is out!


  • The information section was not injected for radicals in the lesson page if the radical was using an image (for example WaniKani / Radical / Morning)


  • You can now mark/unmark items by clicking on their badge circle to turn the badge yellow. Some of the kanji where the tone mark is not visible anymore are already pre-marked. The badge-marked information is stored inside Tampermonkey and should be quite persistent (you can reset it in the options again).

1 Like

Hello! I just found out about this script and immediately download it, but it seems like it isn’t working. I’m sorry if in the end it is just some silliness from my side of the table, but I tried to search lots of kanji and radicals and I couldn’t find anything script-related. I’ve never used any version of this script before the current one (1.1.5), so that may be the cause. I tried all the links in this page too and was still unsuccessful. I’m using Violentmonkey v2.8.20 on Firefox.

I tried a bit to repair it, it seems that Violentmonkey fails to handle unicode properly. They are also unable to provide a button to clear their cache, which shows that the developers are of the same calibre as greasemonkey’s. Finally, loading the files locally work now, using exactly the same files from github fails, no idea why.

Long story short, please switch to tampermonkey for everyones sanity :wink:

Thanks for the fast reply! Besides WaniKani, I don’t really use scripts at all, so have little experience in these matters. I started using violentmonkey after greasymonkey stopped working. Thanks for the tip! I already installed Tampermonkey and your script is working as expected. Sorry for troubling you and, again, thanks for this great script!

@ICerejo I reordered the files a bit, it’s now also working in Violentmonkey, now sure what the problem was. But I think you will be happier in the long run with Tampermonkey :slight_smile:

Latest version is now 1.2.0.

Shouldn’t 磁 be included under the 茲 family?
I just had a lesson for 滋 and 磁 wasn’t included despite having the same reading and tone mark.

1 Like

Thanks! I missed that one, added now.

The script always uses the latest DB file on github, it will show up when your cache expired.

1 Like

Reporting another potential miss.

WaniKani / Kanji / 獲 - this kanji is not listed in the phonetic/semantic section.

WaniKani / Kanji / 護 - this one is apparently of unknown origin but it’s listed under the 蒦 family on pages for 獲 and 護.

Something might be misplaced here but it’s more likely that I’m misunderstanding it.

So you have WaniKani / Radical / Specialty which is listed under the 尃 phonetic family. That’s ok, didn’t even give it a second thought because radicals can differ a bit.
However, I just noticed that WaniKani / Kanji / 専, which is the exact same symbol/kanji, has itself listed as the tone mark - 専.

So basically I am wondering why the radical 専 belongs to 尃 「ふ・はく」and the 専 kanji belongs to 専 「せん」. WHAT HAPPENED HERE.

@gzovak Thanks for your reporting, I fixed the 蒦 part.

For specialty I’m not so sure, I think the problem is with WK :wink:

If you look closely at the “found in kanji” of radical “specialty”, you can see that there is always an additional stroke top right (as in 尃) 博, 薄, 簿, 縛. The kanji “Specialty” itself doesn’t have this stroke (専, previously 專).

So, the radical “specialty” pretends to come from kanji “specialty”, but the kanji where the radical “specialty” appears in are actually using another tone mark (and thus the radical doesn’t completely fit), while the kanji “specialty” is actually also used as a tone mark, but the compounds are quite obscure and not useful. I just chose the “real radical” for specialty arbitrarily.

1 Like

Shouldn’t 業 kanji be included under the 業 phonetic family (even though the tone mark is not useful in this case)?
So far that was usually the case (except the specialty above), so I’m just checking if that’s intentional.

@gzovak Seems like it’s intentional, 菐 ≠ 業, the business radical is not business :wink:

I fixed the second part, I forgot to add 容 → 溶 as a separate tone mark, just added the cross-reference there. Thanks!

@acm2010 There’s one part of the script that I think is misleading and somewhat counterproductive. Sometimes the phonetic component has more than one reading. But at times this makes it look like a phonetic group is more consistent than it actually is and sometimes the phonetic component is downright useless. For example:

In this example we have two kanji that share a phonetic component, but have different readings. And there aren’t any other kanji with this component either. Wouldn’t it make sense to either remove reference to this component or add some feature to indicate that it’s unhelpful?

@seanblue I added all readings for people who also want to go “beyond WK”. While preparing the list I was deleting 80% of the compounds I found because they are jinmeiyou or even more obscure, but for those kanji the tone mark is even more consistent.

For this particular kanji I think I added the readings of its compounds because I couldn’t find the actual reading of the tone mark itself, it is probably only one of the two given. Chinese readings seem to be
1.mí 2.shēn 3.shèn. I can probably kick out the “たん”.

But for most other tone marks there are several more kanji that are not shown for each ON reading.

I’m planning to grey out ON readings that are not appearing in WK (but still show them), and (different topic) handle shifts like しん → じん more gracefully, at the moment it will be marked as no fit, but it is more helpful than nothing.

1 Like

Info incomplete for 辛.