Pitch-Accent Awareness Corner

As evidenced by a couple recent posts, many Japanese learners make it surprisingly far into their studies prior to the shocking discovery that pitch accent exists in Japanese, and that understanding how to acquire it correctly it is essential for speaking Japanese without an obvious accent.

Whether or not to dive into this subject (and when) is apparently slightly controversial, but for those who have decided to take the plunge, here’s a place to share resources, strategies, and questions. I’ll keep a running list of all the resources so please kindly share what you know of!

We do have a couple users with deep knowledge of the subject like @jjatria, and I’m fairly sure Dogen lurks in these forums…


Scripts/Plug-ins

WK Pitch Accent script, created by @Invertex

Anki plug-in for adding pitch-accent info to cards


References with pitch accent info

Apple J-2-J - if you are on a mac, open up the dictionary app and turn on the Japanese to Japanese dictionary and you’ll find one of the clearest references for pitch-accent information available… the information comes from the 大辞林, and you can see the charts here.

Forvo – get recordings of native speakers pronouncing almost any word, and request items they don’t have (I’ve used the request feature and got a pronunciation within a day)

Prosody Tutor Suzuki-Kun

Voicevox
Website (Japanese)
Wanikani Community post

OJAD - Online Japanese Accent Dictionary
http://www.gavo.t.u-tokyo.ac.jp/ojad/eng/search/index – Direct Link to Search Form
http://www.gavo.t.u-tokyo.ac.jp/ojad/ – Home Page

Weblio


Introductions/Overviews

日本語標準アクセントの概要 – a comprehensive overview of the Japanese standard accent.

近世京都・大阪アクセント – an introduction to the pitch-accent for Kansai dialects.

“A quick and dirty overview of Tokyo Pitch Accent” (just a random reddit post but very clearly written – however, please see this post from jjatria for clarifications and corrections!)

Japanese Accent Study Website
(A great little online pitch-accent resource where you can look at lists of words by groups (names, counters, verbs, etc) with pitch-accent indicated. A good introductory resource to catch up on learning pitch accent for words you might already know!)

Video/Audio introduction to Tokyo pitch accent in both English and Japanese (comes from this site).

Speak Japanese Naturally with Fumi-san, a female native speaker on YouTube who provides Pitch Accent practice materials and more.

Campanas de Japanese a Japanese Pitch Accent and Pronunciation YouTube channel by Mei-san, a male native speaker. It also has 3 different speed levels of Amenbo No Uta exercise videos.


Courses

Waseda Pronunciation Course
Japanese Pronunciation for Communication

Dogen
Intro video (first five lessons are free)
Patreon page with the whole curriculum for paid patrons.


Supplemental Help/Exercises

Shadowing Beginner book:

Shadowing Intermediate book:

1日10分の発音練習 (10 minutes a day pronunciation practice) book:

Genki Pitch accent supplement - visual tonal guide for genki dialogues:

82 Likes

Yeah, this is definitely a very overlooked aspect of 日本語, thanks for finding all of these!

2 Likes

But if I sound native they won’t say "えええ 何で日本語分かるの?凄い!上手だよ! every time I say “konnichiwa” with a British accent.

How do I reconcile this?

57 Likes

My friend told me about this website: http://nihongo.hum.tmu.ac.jp/mic-j/home-e.html (it has some other things on pronunciation as well)
Which has this useful introduction to pitch accents: tokyo_0226_outline

I don’t think this is on the list yet, I read it four times, but if it is I apologize xD

(and thank you for making this post!! I’ve been trying to find info about pitch accents for the past month and it’s surprisingly difficult?!)

6 Likes

I find their homepage difficult to navigate. Here is the direct link to their search form if you want to include that instead / as well:

3 Likes

Nice! Now we have a special place of our own :3

Here are the things I think are not entirely accurate:

i. English uses stress.
Stress contains three elements:
A rise in pitch.
A rise in volume.
A slight lengthening of the syllable.

English also uses vowel reduction

English also uses a feature called “vowel reduction”, which basically reduces the phonetic information in non-important vowels, making the accented vowel more prominent by comparison.

Compare the the second /o/ in <photographer> with that in <photograph>. In the first one, in which it is accented, the sound is much more o-like, while in the second, where it is unaccented, it becomes what is caled a schwa, or a central, neutral vowel.

This is important because Japanese does not do vowel reduction, which basically means that vowels in Japanese do not generally change depending on whether or not they are accented.

iii. What is pitch?
It is simply making your voice higher or lower.

Pitch does not technically depend on you

This might be nit-picking, but here goes.

Pitch is the perception of a sound as being higher or lower. Since it is a perceptual feature, it exists only in the brains of those that listen to a sound. It is not an acoustic feature, nor a feature of the sound: it is a feature of the perceived sound.

What is a feature of the sound is its frequency, and frequency is correlated with pitch. But pitch depends on the ability of the listener to perceive those changes as significant differences. A sound can be produced with a higher frequency and still be perceived to have the same pitch.

Think of people that are “tone-deaf”: they can hear two musical notes that are acoustically, factually different, but they will have difficulty identifying them as having a different pitch (because for them, they won’t).

iv. Most important rules

  1. Japanese has two pitches tones: high and low.
  2. The first two syllables mora of a word must be different pitches tones.
  3. [See below]
  4. There can only be one pitch drop in a word. (I.e. if a low syllable mora comes after a high syllable mora, then there cannot be another high syllable tone in the word.)

3. The default pitch pattern for words is for the first syllable to be low and the rest to be high.

Default accent position

This changes depending on the historical origin of the word. It is true that most やまと言葉 are unaccented (low and then high, see photo), but the vast majority of loan words (words that have been borrowed from another language) are indeed accented.

For words that are accented, the default position is for them to be accented on the third mora from the end.

This is from my dissertation, and includes much more detail about this:

Just remember that a word with no arrow starts low.

An easier(?) rule-based mnemonic

An easier way to remember this is that the accent is the drop. If the word is unaccented, there is no drop. So the entire word must be high (because it never falls). But the first and second mora must be different, so the first one has to be low (because the second, and all the others, are high). So you get LHHH…

vii. What counts as a syllable mora in terms of accent marking?

  1. Obvious stuff: ka ki sa shi su tsu etc.
  2. N is always counted as its own syllable mora.
  3. Kyo, kyu, jyu etc. are one syllable mora.
  4. Long vowels. (I.e. Kyuu = 2 syllables moras kyu + u.)

[…]

(Technically these “syllables” are actually called morae, but people are generally more familiar with the word syllable, so I went with that for this 2 minute overview.)

Syllables and moras

This is a pretty big difference, so I don’t think it is useful to “dumb it down”.

With notable exceptions, the Japanese accent system (and certainly the accent system of standard Tokyo Japanese) is not syllable-based it is mora-based. This is also true for traditional Japanese metric: haiku count 5-7-5 moras, not syllables.

A mora is not “a special kind of syllable”. They are entirely different things. The best definition I’ve found for a mora is that a mora is “something of which a long syllable consists of two and a short syllable of one”. It is a metric unit that sits between individual sounds and syllables.

In the case of Japanese, this is easy to remember because the kana alphabets are not syllabic, but moraic: anything that has its own kana (or kana digraph) is a mora.

This includes the katakana vowel lengthener 「ー」, the long vowels, the 「ん」, the 「っ」, etc. All of those are individual moras, and in traditional Japanese metric they should all have the same duration.

So, when you read that the default accented pattern is “three from the back” (「後ろから三番目」), those are always going to be mora.

10 Likes

Can you give an advice on how to make this work with {{type:Field}}? I want to type in beyond Kana reading, because Kana isn’t enough for reading.

1 Like

sorry I’m not totally clear what you’re asking – are you not able to make the plug-in work at all or is there a specific thing you are trying to do that isn’t working?

For this plug-in to work at all:

  • The Japanese Support plug-in must be installed
  • Your note type must have the word “Japanese” in it.
  • Your note must have a field called “Expression” (the content of this field is what the plug-in will look up) and another called “Pronunciation” (this is where it will put the accent pattern).

If those are set up and the pitch-accent plug-in is installed, then if you type a word that the plug-in recognizes into the “expression” field and then press tab, the plug-in should automatically put the pitch-accent pattern into the “pronunciation” field.

1 Like

this sort of blew my mind – I have a friend who wrote a textbook for composers about writing for percussion instruments, and in the process of editing it we spent a year arguing about pitch terminology. In the percussion world, many instruments are commonly referred to as “un-pitched” when the reality is that they actually just have pitch which is unclear/hard to identify, or not notated for various reasons. We wanted to throw out this term because we felt it was misleading ("but… that woodblock isn’t un-pitched… it’s a D#!)

The concept that pitch is a psychological phenomenon whereas frequency is a measurable scientific fact never entered the discussion – we always saw the two terms as just different ways of describing the exact same thing. That may have been a mistake!

So besides human tone-deafness, are there other situations that could cause one person to hear an upstep and another to actually hear a downstep in language?

In music certain instruments with lots of upper partial content can exhibit this: play a low C and a high C on a toy piano and some people will hear an upstep while others will hear a downstep. And if you’ve ever tried to transcribe IDM or music with lots of layered synths, you’ll know how frustrating this phenomenon can be.

Related, here is 10 hours of Shepard Tones:

6 Likes

For those living in Kansai, or who are interested in Kansai, Kyoto, or Osaka dialects, here is a resource I found detailing the pitch patterns of some words:

http://www.akenotsuki.com/kyookotoba/zhisho/

Not exhaustive by any means but interesting to compare with the more standard Tokyo patterns.

Edit: Ah sorry! Just realised it’s been posted at the top of the page.

2 Likes

Thanks. Thanks to you, I managed to make it work, but…

It does mean that it works, though; with the 3 restraints you mentioned.

1 Like

hahaha excellent choice of example word.

I’m not sure why you’re getting those background colors, and the overline isn’t being placed correctly.

When I put the same word in, the card looks like below:

I’m really bad with Anki coding… perhaps @hinekidori can help?

1 Like

that was me taking it from your post, I just did it rul quick :slight_smile: thanks for sharing!

1 Like

This is the effect of {{type:Field}}. I managed to find a tutorial in Anki Manual.

1 Like

ah! I see now, I wasn’t familiar with that feature.

1 Like

That would be quite dramatic. I’ve never heard of language differences that would make some people hear an upstep and others hear a downstep. But it is relatively common to see people that do not hear a difference where others do, and vice-versa.

There are also levels at which you can approach this:

One is simply being able to tell whether two sounds (which might not even be speech sounds) are different, and that is related to the notion of the JND, or Just Noticeable Difference: the minimum physical difference that is actually perceived as such by the brain. In intonation, this is most commonly reported to be between 1.5 and 2 semitones. This would be a phonetic difference.

Another is whether the differences you hear are interpreted as meaningful, or linguistically relevant (in other words, a phonological difference). This question has a lot of applications also for second language learning, and there’s been a lot of studies on cross-linguistic perception. Basically: how difficult is it for speakers of A to acquire the phonological distinctions of B? And why?

The most famous are probably studies done with French and Spanish speakers. In French the position of the stress is neutralised (= it doesn’t change), but in Spanish it is a fundamental part of the language. So the question was whether French native speakers would be able to hear those differences. The initial study found that they were terrible at it, and proposed the notion of “stress deafness”. But this has been softened quite a bit since then, as people have realised that it is not so clear cut (French speakers can learn to perceive those differences, and some of them do not need to be taught at all).

4 Likes

Anyway, I have just finished adding Pitch Accent to about 20k vocabs (half is Core 10k, half is WaniKani, the remaining is my custom vocab lists.). This should be made the official Core 10k.

The add-on seems to have a bug too. If this could be made official, shared users of the deck wouldn’t need to install the add-on.

Also,

Capital E and capital P, actually.

2 Likes

Sorry for the intrusion, but: why? I mean, is the goal only to mask your accent and it has no meaningful difference?

1 Like

Is improving your ability to be easily understood by native speakers an unworthy goal? The closer you get to natural Japanese, the smoother communication becomes. There are meaningful differences between homonyms that are distinguished by pitch, and even if context can solve it, it is distracting to a listener if your intonation is all over the place.

15 Likes

whoops! good catch, thank you.

1 Like