Htrk's Study Log (Pitch Accent and Phonetics)

So, I figured I’d start a log to document some of the progress and challenges I find as I continue to try to build a study routine that accounts for pitch-accent and phonetics.

Problems I’ve Noticed in my Accent

I am a native English speaker with a general American mid-west accent. Problems I’ve noticed (and have started working on) since studying pitch are these:

  • I put accent onto Japanese words in a pattern consistent with English (stress on second or third to last syllable). So, something like 日本語の勉強 would sound like にほんーーごのべんーーきょう as opposed to the correct にほんごのべんきょうーーーーーーーーー in the way I was speaking.

  • Because English stress also extends vowel length on stressed syllables, this also seems to lead to an effect where my accented vowels are too long and the non-accented one’s get clipped.

  • English tends to use the schwa vowel (the “uh” sound in cut) a lot for unstressed vowels, and even often replaces vowels with it (think of how “what are you going to do” can become “whacha gonna do” in casual speech). This is even seen in how English assimilates Japanese loanwords: 漫画 [まんがーー] becomes MAN-guh. 芸者 [げいしゃーーー] becomes GEI-shuh. Paying closer attention, I can also see this tendency leaking into my Japanese sometimes, particularly during rapid speech, making my vowels sloppy.

  • English is syllable based while standard Japanese is mora-based. Thus, Japanese has many mid-syllable pitch shifts (such as in the last syllable in 日本人 [にほんじーーーん]), that I just wasn’t making when I was speaking.

  • Mouth positions for some consonants (such as the aforementioned ワ) simply wasn’t quite right. A good example is the ふ sound. I knew ふ wasn’t like English F, where the top front teeth ride up on the bottom lip when saying it, but I was definitely saying words like フェリー and カフェ with that English F instead of the proper Japanese F sound (the voiceless bilabial fricative, ɸ).

Current Study Progress

I finished reading 日本語アクセント入門, which provided much information about the different patterns and tendencies of pitch accent for different parts of speech (such as verb and adjective conjugation patterns, compound nouns and loanwords) and provided comparisons of the basics between various dialects. It also went a little into how phonology can impact pitch accent placement. I’ve also read supplements from the NHK Accent Dictionary on more of the patterns for different types of nouns. However, it’s a lot of information to internalize and I’ll still need some time to thoroughly review it.

In addition, I’ve made an Anki deck with pitch accent info for some of the most basic vocabulary. I’m taking it slow, only adding 5 words a day and reviewing by referencing audio from the NHK dictionary and getting advice from my wife.

I’ve also been reading more about Japanese phonology and trying to fine-tune my mouth positions. I’ve had a little progress, I think, but some sounds, in particular the vowel ウ, are not yet convincing to my wife when I say them.

In terms of passive study, I’ve been trying to notice pitch accent more when I listen, but I’m still spotty. I can sometimes hear it if it’s a word I already studied or a common pattern, but it’s still very hit and miss, and so I’m reluctant to apply shadowing yet.

In reading, if I notice a word or phrase I’m confident I know the pitch of, I’ll try to read it aloud that way.

So, I’m going kind of slow. One of the first problems I ran into earlier was that applying constant mental focus to pitch accent as I spoke, listened, and read was incredibly tiring and caused me to burnout. I’m trying to manage it in a way that fits into my life with work while not burning out on it, but I’m really hoping my initial vocab studying in Anki will have a sort of “pump priming” effect that will allow me to begin to hear pitch accent more generally across the board when doing listening. I’m still pretty early into it, but I’ll try to keep this updated with more (hopefully shorter) posts on what I discover and the challenges I face.


Well, it’s been about another week, I so here’s what’s been happening.

Progress has been slow, and I’ve actually retooled my study method. After about a month into my deck, I had come to dread doing Anki reviews everyday (which feels strange to me, as Anki was the one consistent cornerstone of my study for JLPT and Kanken), as the process came to resemble something like this:

  1. See a word in Anki.
  2. Most of the time, correctly remember the accent pattern.
  3. Try to pronounce the word.
  4. My wife tells me I’m wrong.
  5. Listen to audio from the dictionary and strain to try to hear the difference from what I was saying and then try to say it correctly while still being told I’m wrong half the time.

There’s no frustration like logically knowing the correct answer but not being able to say it because of native language interference and the fact that I can’t hear the fine differences in native speech and accurately reproduce them yet. It’s a completely different feeling than studying kanji, where my Kanken study and leisure reading formed a virtuous cycle of learning a kanji and immediately being able to discern it when I saw it in a book. Now, I learn the accent, but leisure listening doesn’t seem to be reinforcing it in the same way.

So, I decided to pause on Anki for a while and am instead shifting to a more focussed listening based study routine. I have both the audiobook and physical book version of マチネの終わりに by one of my favorite authors, Hirano Keiichiro. I’ve read the book before, so I know the content. I listen to small bits of the audio a few times, trying to discern the pitch accent, then using the NHK dictionary mark up the accent in a photocopy of pages from the book that I made. This is followed by more listening and trying to read it aloud myself (my wife has agreed to help with this, too, as I obviously cannot grade how well I’m doing). I’ve only just started this and so far I can’t hear the accent well on the initial listen, but it becomes somewhat clearer with repitition and practice. I’m hoping this can help jumpstart my listening some before heading back to memorizing the accents of the vocabulary.

I’m also been perusing through reports NHK released about accent trends upon releasing the latest version of the dictionary. Those can be downloaded in PDF from the NHK website, here.

One of the interesting things from the one on adjectives and verbs is that there is an increasing trend to pronounce adjectives and verbs in the kifuku pattern (起伏式, where the pitch falls before the end) as opposed to the heiban pattern (平板式, where the pitch remains high).

When it comes to verbs, the report has a list of 44 verbs that had heiban as their primary accent in the 1998 version. However, when a survey was conducted for the 2016 version of the dictionary, of the 44, 27 of them now had more people saying the kifuku pattern was an acceptable accent than heiban, and only 5 of the 44 had less than 50% of respondents saying that the kifuku pattern was acceptable. Basically, this means that of the 44, 50%+ of respondants said that kifuku accents were acceptable for 39 of them. There are apparently some other verbs that did switch from kifuku to heiban, but this seems a lot less common.

In contrast, the accent pattern trends for nouns shows a trend towards saying more and more of them in a heiban pattern. The surveys listed are interesting as you can see the language is in constant flux with people’s speech patterns apparently converging towards these trends, especially with younger speakers.

—Other Stuff—
Another interesting tidbit I learned recently is that the Shangainese dialect of Wu Chinese also uses a pitch accent system, which one can read some about here. It’s syllable based, but interestingly the patterns seem to follow a similar principle to Japanese in that once the pitch falls, it does not rise again, leading to a similar array of accent patterns to what one finds in Japanese (not that there’s any relation between the two).

1 Like