There might be a mistake on 順序 and perhaps on other words as well, although I dont know if this has been discussed already. According to Dogen, じょ and other combination hiragana are 1 mora, but the script gives this on wanikani Screenshot - 19f21f7c9529dac9b7691ccea50a85b5 - Gyazo
This is also nonsensical because the second pitch must be different from the first, and what is written is not even 頭高
The script’s not working at all for me, at least not on the individual vocab pages.
As for the specific word in question, it looks right to me. It’s a stylistic choice for the script to display the dot above the ゅ (etc.). That doesn’t change the fact that it drops after the first mora.
Well I think the whole point is that 1 dot = 1 mora. I’ve realized now that the script does this for all words with combination hiragana.
Unless you use a method to get userscripts to run on mobile, you can’t do this within WaniKani itself, but you can always crosscheck individual words with this userscript’s source, weblio.jp, or any other standard pitch accent resource.
I don’t think the intention was that 1 dot = 1 mora – I think it was put there so you can easily track where the downstep is to the kana below. But in whichever case, counting how many mora there are is mostly irrelevant; the point of the script is to easily recognise where (or if) the downstep occurs in each vocabulary, which is accomplished perfectly fine, in my opinion.
I don’t want to sound too desperate, but I would really appreciate it if someone with the skill to do so could update the script to fix its compatibility in vocab pages. I’m happy it still works at all, but in the near future I would really like to go through WaniKani’s vocabulary starting from the early levels and put them in an Anki deck for the sole purpose of pitch accent study, but doing this by individually looking up every word would be a significantly more tedious task.
WaniKani must have changed the UI a bit, so now the script can’t find where on the UI to put the pitch diagram. Unfortunately the code is a bit of a mess, so I’m not sure what we’d have to change to fix it.
Sorry if this has been asked already, but this is a long post and searching it didn’t find me a definitive answer. I just started using this script today and I would like to know, when the pitch graph does not match with the audio provided by WaniKani, which one should I trust? How do I know which is correct?
The pitch accents in this script are sourced from http://weblio.jp. However, the script does/did this by scraping the website in an automated way and can have mistakes. When you notice a discrepancy, I would recommend going to weblio and verifying the pitch information there. Personally, I’d be inclined to trust the information there more than the audio in WaniKani.
Well, the other caveat with the WK audio is that you also have to be able that trust your ears, and there have been a few times I’ve seen people say they heard a recording one way and then natives I asked disagreed that there was an issue.
It seemed different on at least a few during my review session, but it was like 90 items and I don’t remember specifically now, so I’ll inquire if I notice any like that again next time.
What does the script do if multiple accents are acceptable? Does it display them all?
If a word can always be pronounced using multiple pitch accents it displays all of them. The case that it doesn’t handle properly is when the pitch accent is different based on part of speech. That’s simply because the creator of the script never bothered figuring out how to grab that information from weblio.
Ah, yeah, I can’t think of any off the top of my head from WK, but you have stuff like ふわふわしたケーキ, where ふわふわ has a 1 accent because it’s an adverb there, and then in ふわふわのケーキ it has a 0 accent because it’s a の adjective.
I can’t think of any on WaniKani either, but there definitely are some because I remember seeing them in the past.
I recently started using this script - it’s definitely cool! Thanks for making it.
Most words seems to have the Heiban pattern. What I don’t get is, if it’s supposed to sound
flat, why does it go up? Like, am I supposed to pitch it up or say it flat? Usually the sound samples
say them flat.
I always hear heiban as slightly up. Check this out:
Thanks for the link, even if I can’t really make much of it. I guess if I could read that
I wouldn’t really be using Wanikani.
So heiban isn’t flat then? What difference does it have from the other pattern where it goes up,
頭高 or atamadaka? Sorry I wasn’t able to use the link.
This one should be in English I think: http://www.gavo.t.u-tokyo.ac.jp/ojad/eng/phrasing/index
Basically the difference is that heiban never goes down. It starts low and goes up after the first mora. Then stays high.
The most important rule is that for a given word the pitch must go up or down (i.e. it has to change) after the first mora and that after it has gone down it can’t go back up. (If I remember correctly.)
Ooh - I am supposed to type in Japanese and then I get a line describing the pitch through-out.
Thanks!
Actually in my previous post I was wrong. The pitch pattern that I am confusing heiban with is not atamadaka (cuz atamadaka starts high and goes down), it is odaka. In odaka it starts low and then goes high and stays high. In other words, same as heiban. Unless heiban pitch is not as strong?
Yeah, the pattern where it goes high and then goes low and stays low is nakadaka.
You can’t see/hear the difference when looking at the word in isolation. The difference is that with heiban the word ends high and the following particle is high, while with odaka the following particle is low. (This is all on jisho by the way. 
)