Detailed guide to pronunciation of Japanese
The things no one usually mentions in japanese tutorials
Pronunciation of Japanese may seem simple, but it is, in fact, not exactly. You may have been struggling with pronunciation of Japanese [r] sound or trying to get rid of the stress-timing, and all those things are very good to getting you to around 50-70% of native pronunciation.
If you want a perfect 100%, in addition to reading this post, you would need to dedicate years of studying to pitch accent, which can come either naturally, given that you have a good hearing, or by reading materials regarding details of pitch accent of different words, and the rules of pitch accent for verbs. It’s a large and complicated topic, which I will not cover in this post.
In this article I want to write about the aspects of Japanese pronunciation that can get you from those rookie numbers all the way up to 80-90% of Japanese pronunciation. If you are not interested in learning the right pronunciation, you might still be curious about some aspects of it. In any case, if you’re interested, you’re welcome to read the rest of the article.
Also note that this is an article summarising my personal experience as well as information I have found on internet throughout my path of learning Japanese. Some parts of it may be wrong, and if you have some corrections please send them to me.
1. Plain timing
In English, stressed syllable is pronounced differently from the other syllables: Stressed syllables are often louder than non-stressed syllables, and they may have a higher or lower pitch.
Let’s take the english word “complicated” and try to graph the timing of the syllables as they are pronounced. I have recoded myself saying the word “complicated” and then marked the points where I pronounce different syllables. Note that I am not a native English speaker, so the timings may be a bit wrong, and not to brag, but me and other native speakers consider my pronunciation of english fairly native-like. Anyway, here’s the result:
Notice how the stressed syllable “ca” is pronounced longer than all other symbols. We note that syllable timing in english is uneven.
Japanese syllable timing is not like English, because, first of all it doesn’t have stress. Instead all syllables appear at a constant rate as if read with a metronome. I did the same thing with Japanese word 新しい [atarashii], which has 4 syllables, and here’s the result:
I’m also not native in Japanese, but I’m fairly confident in my pronunciation. I wasn’t 100% sure, so I downloaded an audio of a native speaker pronouncing the word and graphed it too. Here’s the result:
The syllables appear at constant rate.
Training yourself to pronounce syllables at a constant rate is one thing you can do to make your pronunciation sound more native-like.
2. No vowel reduction
Another stress-related difference is that unlike english, Japanese does not have vowel reduction. This doubly hurt for me as russian, where vowel reduction is literally everywhere in the language and I could not remember the words with lots of [a]/[o] vowels and was constantly mispronouncing the word 友達 (tomodachi).
In English unstressed vowels often get reduced to a [ə] (schwa) sound. Examples:
Reduced a, as in about [əˈbaʊ̯t]
Reduced e, as in taken [ˈtʰeɪ̯kən]
Therefore it’s common for English speakers to pronounce Japanese words with a [ə] sound, and Japanese people can in fact detect foreigners based on that. I will talk about the actual sounds for vowels later in this post, but for now just take a note to properly pronounce the vowels in all syllables.
3. The timing of ん
Mora is a linguistic term, referring to units of timing. While English is a stress-timed language, Japanese is a syllable-timed language, meaning that each syllable is pronounced with it’s own unit of time.
This last part is actually a bit incorrect, because not all moras in japanese are actually syllables. ん (n) is actually not a syllable, but it is a mora, that is, it gets its own unit of time, just like other moras.
Fun fact: if you listen to any Japanese song, the timing of words becomes loses it’s even-ness, but each mora is sung according to music’s rhythm, so in terms of rhythm the timing is still plain. In songs, ん can be sung either in its own rhythmic unit of a song, or it can get added to the previous syllable, so sometimes けん is a single syllable composed of 2 moras.
Quick aside: about moras
To continue the discussion I should tell you about the moras. All the moras in Japanese are usually organised into a table like this one:
Sometimes this table is written horizontally, however in Japan people only draw it like this, and the reason for that is how they refer to groups of moras on the table.
This table is organised into rows, which have the “same” starting constant and columns which have the same following vowel. When we refer to moras of the same starting constant, for example na/ni/nu/ne/no, I will say “na-row” (jp. な行). So the ra-row will be referring to moras: ra/ri/ru/re/ro.
You can notice how some starting constants of the same row aren’t actually the same. For example the sa- and ta-rows. Also there are missing entries in the table, particularly in the wa-row and ya-row. Most of those moras do not spell any japanese words, but are still sometimes used, here’s for example the “we” mora: ゑ.
Also this particular table puts ん in the wa-row, in the place of を (wo). I’d use a different table or make my own, but I’m too lazy to spend time on that. ん is usually set aside, not being part of particular row. It is kind of special in that regard.
Not all consonants are reflected on the table:
だ (た with dakuten) - pronounced as da
が (か with dakuten) - pronounced as ga
ば (は with dakuten) - pronounced as ba
ぱ (は with handakuten) - pronounced as pa
The moras with dakuten and handakuten follow the same pattern with a few exceptions:
ぢ (di) - pronounced as ji
づ (du, dzu) - pronounced as dzu, also sometimes as zu
Most consonants are pronounced like in English, with some exceptions, the more detailed guide to pronunciation of the consonants will follow later in this article.
4. Pronunciation of vowels
Japanese has 5 vowels: a/i/u/e/o. Teaching pronunciation of vowels through analogy with English is a bit of a wrong move, because the pronunciation of English vowels in words can differ by dialects and region. But that way is easy, so i’ll leave it in, however if you feel confused google the IPA symbol for the vowel and open wikipedia page. It usually has an audio recording of a person pronouncing the particular sound.
The vowel “a” is pronounced like the “a” in the word “bra” (except Canadian and Californian english). The IPA symbol for the sound is [ä].
The vowel “i” is pronounced like “i” in the word “bit” (except Australian English), or “ee” in the word “free”, though this word usually is pronounced with a diphthong. The IPA for this sound is [i].
The vowel “u” is not met in English except in Californian English (“goose”). The IPA symbol is ɯ̟.
The vowel “e” is pronounced like in the word “let”. The IPA symbol is [e̞].
The vowel “o” is pronounced like in the word “thought”. The IPA symbol is [o̞].
The other note about vowels is that whenever “e” is followed by an “i” it’s pronounced like the long “e” sound. Similarly when “o” is followed by an “u” it’s pronounced like the long “o” sound. For example the word “keigo” would be pronounced like [ke:go].
And also it’s worth noting that some “i”/”u” syllables can be devoiced. The word 靴 (“kutsu”, meaning: shoe) has the first syllable devoiced. It sounds like “ktsu” when said out loud. 季節 (kisetsu, meaning: season) sounds like “kʲsets”.
6. The Japanese “r”
It’s one of those consonants that is very different from English “r”, which makes it a little bit hard to learn for English speakers. People describe this as a mix of “r” and “l”, but this is a bit misleading.
The truth is actually a bit closer. It’s pronounced the same as “tt” in the word “better” American English or “t” in the word “water”. For details I highly recommend watching Dogen’s video guide to pronouncing the Japanese “r” sound.
Note that sometimes “r” is pronounced like “l”. There’s probably no rule, they are mostly interchangeable.
7. The ふ (“fu”) pronunciation
The sound “f” in Japanese is unlike English “f” sound, which is pronounced using the top lip and the bottom teeth. Japanese “f” is a bilabial sound, produced with both lips. That said, the “fu” sound can be pronounced as “hu”, but it’s more like something in between the “fu” and the “hu”. This handwavy explanation is unfortunately the best I can give. If you’re looking to improve your japanese try imitating native speakers :D
8. The “g” pronunciation
The ga-row consonants can be pronounced like in english, however there’s also a pronunciation of this as the “ng” sound (Like “ng” in the word “doing”). It may feel weird at first, because in english there are no words beginning with this sound. Native speakers use both pronunciations however I feel worth it pointing out this fact.
9. The “w” pronunciation
The English “w” is pronounced by rounding your lips and opening them up. To pronounce Japanese “w” do the same but do not round the lips.
10. The various pronunciations of ん
The syllable ん is a very special one in many aspects, including the aspect of pronunciation. Depending on the consonant following it, it can be pronounced in different ways:
Before “m”, “b”, “p” - pronounced like English “m”
Before “k”, “g” - pronounced like English “ng”
Before “n”, “r”, “t”, “d” - pronounced like English “n”
Before “nya”, “nyo”, “nyu” - pronounced like spanish ñ, or n in the word “new” (British English)
Before vowels - kind of like saying g without closing the throat, just like when like saying w. IPA for the sound is [ɰ̃]
11. Interchangeability of ず/づ
Sometimes the words with づ (dzu) are pronounced with ず (zu), like the word つづく can be pronounced tsuzuku. Similarly some words spelled with ず are pronounced with づ. Keeping this in mind can help you in studying Japanese.
12. Pronunciation of ぢ (di)
Pronounced like ji.