Curiously enough, Hepburn romanization fixes some ambiguities in Japanese (Japanese written in kana alone) while introducing others.
The ō in Hepburn could correspond to おう or おお or オー. That's an ambiguity.
Where does Hepburn disambiguate?
In Japanese, an E column kana followed by I sometimes makes a long E, like in 先生 (sen + sei -> sensē). The "SEI" is one unit. But in other situations it does not, like in a compound word ending in the E kana, where the second word starts with I. For instance 酒色 (sake + iro -> sakeiro, not sakēro).
Hepburn distinguishes these; the hiragana spelling does not!
This is one of the issues that makes it very hard to read Japanese that is written with hiragana only, rather than kanji. No word breaks and not knowing whether せい is supposed to be sē or sei.
There are curiosities like karaage which is "kara" (crust) + "age" (fried thing). A lot of the time it is pronounced as karāge, because of the way RA and A come together. Other times you hear a kind of flutter in it which articulates two A's.
I have no idea which romanization to use. Flip a coin?
> The kara comes from a country name and refers to a style of cooking
My understanding is that the exact etymology is unknown. It's often written with the letter that references the tang dynasty, but the thing is there's no particular reason to think the Chinese introduced the style of cooking to Japan - although it is true that there was such a thing as fried chicken in 7th century China!
Another kanji-ization of the word uses the kara from karate (meaning air or empty, in karate it's "empty hand") and I find this equally plausible as karaage is fried with a very small amount of batter ("in air").
Either way they're both essentially competing "kanji backronyms" seeking to retcon an existing word as spoken; there's no real right or wrong answer.
I somehow always keep forgetting that the kara part is that kanji that looks like the one for sugar without the kome hen.
Still, that sort of thing in general still leaves room for it having been word play. Like tempura being originally from Portuguese, having nothing to do with 天.
Japanese spelling often plays gaslighting head games.
> In Japanese, an E column kana followed by I sometimes makes a long E, like in 先生 (sen + sei -> sensē).
While it is sometimes difficult to discern the combined E and I sound, especially for non-native speakers, the word 先生 (sensei) is technically pronounced "sensei" and should be spelled that way to distinguish it from words with long E sounds, such as ええ (ee) and お姉さん (oneesan). Similarly, the OU in 東京 (toukyou) and the OO in 大きな (ookina) are different and should be spelled differently. I hope this helps.
Sure, and in a Japanese song, "sensei" can yield four beats or notes SE/N/SE/I.
But spelling out and singing aren't normal speech. Spelling/singing can break apart diphthongs, like NAI becomes NA-I.
生 is not written with い due to the /e:/ having a different sound from that one in from おねえさん. It does not (when you aren't spelling). It is written the way it is for ancient historic reasons.
> Similarly, the OU in 東京 (toukyou) and the OO in 大きな (ookina) are different
In the phonetic alphabet it's /e:/ vs. /ei/ and /o:/ vs. /ou/.
If you're an English speaker, you can be forgiven for a very stereotypical trait of the English accent. English speakers have a real hard time with the /e/ or /e:/ sounds as well as the /o/ and /o:/ sounds. Most English dialects don't have either a monophthong /e/ or /o/. Both the long and short tend to get heard as /eɪ/ and /oʊ/.
French enchanté /ɑ̃ ʃɑ̃ te/ is heard and borrowed as /ɑn.ʃɑn.teɪ/. German gehen /ge:n/ is heard as "gain" /geɪn/. And Japanese /o:/ and /ou/ both get heard as /oʊ/.
It's arguably a minimal pair in Japanese: 負う /ou/ (to carry), 王 /o:/ (king).
The main issues probably arise on official documents and stuff with financial impact.
Like how many people end up with the same romanized name while being distinct in other alphabets. Then discrepancies between the different systems because they usually are sloppy on the handling of these matters.
Now that most stuff is electronic, these small differences can have wider effects and be a PITA to fix.
They’re not the same. おう is discernible from おお, and the difference can be important.
That said, this is far from the most important problem in Japanese pronunciation for westerners, and at speed the distinction between them can become very subtle.
You need to know previously the word to write from Hepburn to Kana when "ō" is present because data is lost in such transliteration from おう or おお or オー to Hepburn.
The internet is full of romanji written incorrectly with "o" alone when it should be "ou" or "oo" due "ō" ASCII conversion errors at one moment.
(The sooner a beginner embrace Hiragana and Katakana, the better)
What's interesting is that they address this problem where the latin alphabet introduces the ambiguity (Is genin げんいん or げにん? Hepburn goes with gen'in for the former to avoid ambiguity), so they could have extended that to sake'iro and applied the same strategy when the ambiguity comes from kana itself.
For what it’s worth as a long time learner of Japanese, none of these ambiguities has ever confused me nor hindered my ability to be perceived as natural to native speakers, so I think that this ambiguity is not such a big deal.
To me, Hepburn’s strength relative to the old government romanization is that it increases the likelihood that an English speaker will make approximately the right sound when reading some Romaji, and that people seem to prefer it in general.
Transcription gets even messier when more than two languages are involved. Russian uses the Polianov system as a "cyrillization" method. It's neither Hepburn nor Kunrei-shiki, which can be confusing if you are a Russian Language learner and know Japanese or English.
Some Japanese words entered Russian not directly, but through English. In these cases, the word is first romanized using Hepburn, and then adapted to Russian using English-to-Russian rules. A classic example is 寿司, which Polianov would render as суси (susi), but Russians mostly know as суши (sushi). Then there are words which actually do faithfully follow Polianov, as in 新宿, which is written as Синдзуку (Sindzuku) instead of Шинджуку (Shinjuku).
Another example of JP→EN→RU is Nintendo's character Yoshi:
By Polivanov, it should have become "Ёси" but since it came to RU via EN, it is written as "Йоши".
しんじゅく (Cиндзюку, Sindzyuku) is an interesting case, as it has both し and じゅ in it. This is where Polivanov is similar to Kunrei. OTOH, Fukushima is cyrillized as Фукусима (Fukusima), where the ふ is a fu in Hepburn, hu in Kunrei and fu in Polivanov but し is not shi as in Hepburn, but si as in Kunrei.
The language school I attended all but banned romanization. The idea was to learn, practice, and finally internalize kana and kanji as quickly as possible. Hepburn is just a band-aid when it comes to language study.
For people not interested in learning Japanese, however, a unified romanization could have its benefits. It just never struck me as particularly inconsistent to begin with, even after so many years living there.
There’s another school of teaching, where kana and kanji are banned for the first 2-3 semesters because they are a distraction to learn and internalize words and grammar.
I’ve met a few students of this textbook system when I was on exchange and my impression was that they were very skilled at Japanese for the amount of time they’ve been a student and what they told about their seniors was they pick up kanji fast, since they already know the words.
The big problem of course is that it is completely incompatible with other schools. Where do you place them when they go on exchange? With the n3 or n5 students?
Anyway, I always thought it was interesting that the exact antithesis of RTK* exists and works.
*RTK or “remembering the kanji” is a system that teaches all kanji before student learn their first word. It’s quite popular online as it lends itself very well to solo studying.
> *RTK or “remembering the kanji” is a system that teaches all kanji before student learn their first word. It’s quite popular online as it lends itself very well to solo studying.
One thing I have found over the years, I have never met a foreigner living in Japan who has used it extensively. (Many were aware of it, but few used it heavily.) However, there is a lively community of online learners who use it. (Don't read that as a judgement against using it; this is simply an observation.)
I was surprised to read this part:
> a system that teaches all kanji before student learn their first word
I have never heard this description before. I always thought it was a learning aid to use mnemonics to remember the meaning of individual kanji. If someone can complete all volumes of RTK before "learn[ing] their first word", I would be stunned. It would be a feat of super-human level of memorization and recall. That said, the Internet is a huge place with billions of people. There will be somebody, somewhere who took this path and is happy to tell you about their success using it.
There’s another school of teaching, which bans all reading, writing, and speaking altogether in favor of exclusively native speaker verbal input for the first 6-12+ months of learning. Some YouTubers seem to like the idea of this, though sounds pretty extreme.
I have always felt furigana bridges that gap well enough in written learning. The downside is that it might become a crutch, but it can't for long if you are serious about learning reading. Native materials pretty quickly drop furigana.
Like with a lot of things like this, if you learn for long enough the differences in the major approaches work themselves out.
Kunrei-shiki is intended for domestic Japanese use. That's why it results in spellings that don't make logical sense for any Latin-based phonology. It's too focused on round trip unambiguity at the cost of phonetic clarity for non-Japanese. My big peeve is the company Mitutoyo using K-S, which everyone mispronounces because they don't know it's a poor transcription of "Mitsutoyo".
Yeah my impression was the Orthography is pretty consistent compared to English.
From what I understand this isn't the first time they've made some kind of change to orthography, I remember reading something about updating offical use of certain kana to reflect more modern pronunciations. It wasn't a dramatic change.
It's interesting to see some countries just have this centralised influence over something like how their language is written as they're the main ones speaking it, as opposed to English.
> Yeah my impression was the Orthography is pretty consistent compared to English.
As a native English speaker, I have learned this watching non-natives try to learn English spelling over the years. It is hell! I studied French in middle school and high school. I remember there being a similar level of ambiguity in their orthography (similar to English).
One weird thing that I have noticed when Japanese native speakers write emails in English: Why don't they use basic spell check? I'm talking about stuff as basic as: "teh" -> "the". Spell checkers from the early 1990s could easily correct these issues. To be clear, I rarely have an issue to understand the meaning of their emails (as a native speaker, it is very easy to skip over minor spelling and grammar mistakes), but I wonder: Why not spell check before you send?
Hepburn is poorly supported in some input methods, like on Windows. If you want to type kōen or whatever, you really have to work for that ō. It's better now on mobile devices and MacOS (what I'm using now): I just long-pressed o and picked ō from a pop-up.
That's one aspect I really love about macOS. I'm from a small country so nearly no one makes hardware with our exact layout, but with macOS I can always just long press to fill in the gaps. I just wish all apps used native inputs, not some weird half-baked solution they built themselves.
I rarely miss Linux, but I liked being able to have compose keys, most of which were very logical and fast to type. Now on MacOS, I either have to know the option (alt) combination or long press, which makes my writing with accents way slower.
What's the best way to type Japanese on Windows? (I have a QWERTY keyboard)
On mobile I just switch to the hiragana keyboard, but that obviously isn't a sane option on desktop unless I'm clicking all the characters with a mouse?
This is a good question. I have seen a wide variety over the years from native Japanese speakers. Some use the 1990s-style kana keyboard. Some use romaji input where real-time software (called an IME) automatically suggests conversion to the final Japanese word (katakana/hiragana/kanji, etc.). On a mobile phone there is usually an option to do 1990s feature phone style kana input, where the 12 key input is shown, and you press one key as many times as necessary to rotate to the correct kana that you wish to input. You can see young girls with frighteningly long fingers nails jamming away -- chatting with their friends via mobile text (Line, SMS, etc.). Their "touch memory" (and sensitivity) must be jaw-droppingly good -- like a professional drummer or something similar.
Native Cantonese speakers in Hongkong have similarly diverse input methods. I've even seen tiny digital draw pads at the public library. It is pretty exciting (to me!) to watch an elderly person furiously scribbling away on the pad, inputting traditional Chinese charaters to search something on the Internet or in the media catalog. I think it is very cool that public library makes a strong effort to empower all types of users.
Using the example from the top-level comment, you would install an IME, switch to hiragana mode, start typing "kouen" and convert to kanji when you see the right suggestion.
It might sound complicated at first, but you can do it pretty fast once you get used to it.
MS-IME or Google Japanese Input. (whatever)-Mozc on Linux. Use "IME On" mode for Japanese, "IME Off" mode for alphanumeric text inbetween.
"shio ha natoriumu[Space][Return][ImeOff](Na)[ImeOn] to enso[Space][Return][ImeOff](Cl)[ImeOn] kara dekite imasu[Return]"
-> "しおはなとりうむ(Na)とえんそ(Cl)からできています"
-> "塩はナトリウム(Na)と塩素(Cl)からできています"
(NOTE: spaces added for legibility)
Most Japanese users use this "romaji" input - which is more vibe heuristics based and not highly consistent with existing romanizations hence the change. Some use "kana" with full 51 Hiragana symbols on JIS keyboard(with ろ/backslash/underscore key to left of RShift, which makes it incompatible with ISO). I think "most people don't do this anymore" remarks refer to the fact that everyone's on the phone, and uses the "flick" input.
When it comes to input "best" is highly subjective, but with that said: Just adding Japanese support in the system language settings is fine.
Standard Qwerty keyboards are well supported, you'll need to either check the key shortcut to switch between inputs or do it with the mouse if it's infrequent enough.
People using it daily will tweak a lot more, have a straight to IME and straight out of IME key instead of the default switching pattern, potentially add more tweaks to always have half-width space and ponctuation whatever the mode they're in etc., but that's a rabbit-hole you'll be free to fall into.
BTW the reverse works well enough: Windows has a specific mode to force US ANSI on JIS layouts and still use the additional japanese keys. Kinda fun they felt the need to leave that mode in.
As others have said, people prefer different ways. My wife (Japanese) writes on Windows (Japanese edition) in romaji, and she's very fast. But she also says that in fact most Japanese (at least of her generation) don't write that way (they presumably use those small kana letters on Japanese-variant keyboards). As a non-native I also write the way she does, though I'm on Linux. I'm not sure why my wife writes using romaji, I should ask.. she wasn't an English speaker or anything, so why that worked for her I don't know.
I don't know now, but for the longest time, Google made a much better Japanese IME for Windows than Microsoft ("Google Japanese Input"). I started using it when running into reliability issues, like disappearing kanji dictionary, or frozen switching between roman and hiragana.
Assuming Microsoft's Japanese IME is still a dumpster fire, and the Google one has not succumbed to Googleshitification, that would be a way to go.
To enable the Microsoft IME there are some rituals to go through like adding the Japanese language and then a Japanese keyboard under that. It will download some materials, like fonts and dictionaries. A reboot is typically not required, I think, unless you make Japanese the primary language.
Once you have the keyboard, LeftShift + LeftAlt chord goes among the input methods. Ctrl + CapsLock toggles hiragana/romaji input. I think these are the same for Google or MS input.
Note: bitwize is talking about how to do it on Linux. Which is the best way in my biased opinion. Perhaps not the best mapping for people who use it regularly but is awesome for those who use it irregularly. We can usually guess how to do weird diacritics without having to look it up.
The article says the new style says that you can use either a macron or a doubled letter, but it's not clear if that's supported for keyboard input on various platforms.
(The bugs I've experienced: it doesn't properly disable itself during video games, despite claiming to do so; sometimes the popup seem to come up when I swear I didn't press the shortcut keys; rarely, the popup gets stuck on screen and needs to be Alt+F4'ed.)
The old official system arguably makes more sense from a Japanese perspective.
If you look at the kana, the Japanese syllabic writing system, they have this ordering: ka ki ku ke ko, sa shi su se so, ta chitsu te to, etc. If you follow the regularity where there should be a "ti" sound there is no "ti" sound and it happens to be pronounced "chi".
One common analysis holds that the underlying phonemes really are: ta ti tu te to. Traditional Japanese grammarians usually analyzed it this way. And they were historically pronounced that way: it has arisen out of relatively recent sound change. Somewhat like how some British speakers pronounce "Tuesday" such that it sounds much like "Chews-day" to speakers of other dialects. Affrication in a fixed context. The t phoneme triggers that kind of affrication obligatorily in Japanese, before the i vowel or y glide.
Some disagree with this as overly theoretic and based excessively on historical linguistics, and they insist that sh and f and ch are distinct phonemes in Japanese. But the Japanese writing system itself treats it as if they were not.
If you are learning Japanese it makes sense to pick a system that reflects the internal logic of kana spelling. If you want to just approximately pronounce Japanese words in English then you want something that reflects the logic of English spelling.
These two goals are always in tension. Mandarin pinyin, for example, was designed to reflect the logic of Mandarin phonology in a consistent way. It's not meant to be easily pronounceable by English speakers. It's to enable Mandarin speakers to look up words in a dictionary or for students of the language to study Mandarin. Though it has ended up used as a pronunciation guide for English speakers. And that often doesn't go well; a lot of English speakers don't know what to do with the q's and x's.
It's a change in purpose. Nihon-shiki was invented to teach Japanese people the Latin alphabet, with a view to replacing kana/kanji with the Latin alphabet. Therefore being understandable to someone with a good idea of the kana layout was the priority.
Hepburn was designed to teach non-Japanese people Japanese, therefore matching well to European (especially English) sounds was considered more important.
Suggesting Japanese romanise is a fringe position these days, much much more so than in the 1880s or the immediate aftermath of WW2, and making that kind of change is much easier when you have a population going from illiterate to literate than in a modern society, so nobody's seriously considered Nihon-shiki (or its slightly modernised descendent, Kunrei-shiki) a gateway to romanising Japanese for the Japanese for a long time now.
So this is sort of an official recognition that the primary purpose of romaji is for the benefit of foreigners.
This is the same reason why I'm disappointed that Pinyin won over Wade-Giles. If Hepburn can be acknowledged to be better than Kurei-Shiki, then Wade-Giles is also better than Pinyin. At the very least we'll no longer have to deal with words containing q that's pronounced nowhere near q. Although admittedly it does produce some exotic looking words and boon for Scrabble players.
I don't know the details history of the system's development, however I notice that with Kunrei everything spelling is neatly 2 characters while with Hepburn it may be 2 or 3 characters:
Kunrei: ki si ti ni hi mi
Hepburn: ki shi chi ni hi mi
The politics of the issue is obviously that Hepburn is older and an American system while Nihon and Kunrei are very purposely domestic (Nihon "is much more regular than Hepburn romanization, and unlike Hepburn's system, it makes no effort to make itself easier to pronounce for English-speakers" [1]). Apparently, Hepburn was later imposed by US occupying forces in 1945.
Perhaps 80 years is long enough and suitable to effect the change officially with no loss of face.
> It sounds way closer to the spoken sounds, at least to my western ears.
That's the thing... to some other non-English language speakers, the existing/old romanization method actually is more accurate regarding how the letters would be pronounced to them, especially coming from languages that don't have the same e.g. [ch] or [ts] sounds as written with Hepburn.
The one technical downside I would say to this change is, 1:1 machine transliteration is no longer possible with Hepburn.
One issue holding back the adoption of Hepburn has been that the standard national curriculum (gakushū shidō yōryō) calls for all children to be taught romaji beginning in the third grade (previously fourth grade) of elementary school. It's taught in Kokugo (national language, i.e., Japanese) classes and included in those textbooks, as romaji characters are used in Japanese alongside kana and kanji as well as, increasingly, in daily life (user names, passwords, etc.). At that age, native speakers of Japanese can acquire kunreishiki more easily, as the consonant representation corresponds more closely to the Japanese phonology that they have internalized.
For pinyin representation of Mandarin, these are very different sounds, while the equivalent (identical) Mandarin pinyin representation of し, じ, つ would be xi, ji, cu. I'm not as familiar with romanization systems closer to Latin pronunciations, but for Wade Giles it would probably be written like shi, chi, tsu.
I didn't really have a problem with the spoken sounds when learning in school - we were also required to take Maori lessons as well and a lot of the sounds are shared. Quite interesting really.
Edit: actually the only one I ever had an issue with was one of my homestay's names "Ryouhei"...that Ryou sound...it's like the Y stops me from rolling the R properly, so odd.
You mean, if you would apply the inverse of the standard romanization of Mandarin, the resulting sound would be closer to the Japanese sound, if starting from the Kunrei spelling than if starting from the Hepburn spelling?
About a decade ago, I became a fan of the remarkable Japanese child prodigy drummer Kanade Sato. That lead to me to learn the surprising fact that Japan has 4 writing systems: kanji, hiragana, katakana, and romanji.
Here's the video that got me interested in Sato www.youtube.com/watch?v=XYpFL08m5fQ&list=RDXYpFL08m5fQ&start_radio=1
Is been 25 years since I took Japanese in highschool but I'm relatively certain that our textbooks had ち romanized as tchi which from my recollection seems more accurate to its actual common pronunciation.
Perhaps only in the case where it's preceded by the small tsu? E.g. "一人ぼっち" -> "hitori bo[tsu]chi" -> "hitori botchi"? That's what Wikipedia says [1], although I think it's also common to (incorrectly?) use "bocchi" instead.
I live in Thailand and I cannot get over the fact that romanization is (seemingly?) completely unstandardized. Even government signage uses different English spelling of Thai words.
You should have seen Taiwan in the 1990s. It was a hot mess of older Western romanization systems, historical and dialectical exceptions, competing Taiwanese and pro-China sensibilities, a widely used international standard (pinyin), and lots of confusion in official and private circles about the proper way to write names and locations using the Latin alphabet. In 1998, the City of Taipei even made up its own Romanization system for street names at the behest of its then-new mayor, a supporter of Taiwan independence (https://pinyin.info/news/2019/article-on-early-tongyong-piny...).
The chart halfway down this blog post lays out some of the challenges once the hanyu pinyin standard was instituted in 2009:
So that’s why people in Taiwan can’t spell anything consistently and why all the English-language newspapers spell the same things differently. As for me, I’m giving up on trying to remember how everyone spells their name. I know lots of people, especially Taiwan nationalists, dislike having the PRC hanyu pinyin system. I dislike imposing it upon them. However, in only three weeks, I’ve found myself spelling the same thing in multiple ways and wasting time looking up how I did it last time. Since almost no one reads my blog anyway, I’ll do it the way that’s most convenient for me.
I’ll also always provide the Chinese characters so that people who can read them know who I’m talking about.
Yeah, my full names are Jeremia Josiah, and on my work permit they wrote the Thai version as เจอเรเมีย โยชิอา. I cannot figure out why they chose to use จ for the J in Jeremia but ย for the J in Josiah. Both are pronounced the same and I would consider จ the correct choice. I would consider ย more correct for representing a word with Y.
Korea is stuck in a funny middle ground, where names like cities or railway stations all follow the standard without exception, while personal or corporate names are in a state of total chaos. So the cell phone maker is Samsung, but the subway station in Seoul is Samseong, even though they're written and pronounced in the same way in Korean. (No, they aren't related.)
It's unfortunate but I don't think it'll get fixed any time soon. Nobody wants to be called Mr. I, O, U, An, or No. (The most common romanization for these family names would be: Lee, Oh, Woo, Ahn, and Roh.)
No country is going to force their big multinationals to change their international name they chose back in the 50s and are now known as world-wide. Personal names aren't too chaotic either, as the choice presented when choosing a romanization is limited, people can't just make stuff up on the ground. They're off, but generally in the same ways.
> Nobody wants to be called Mr. I, O, U, An, or No.
An is pretty common - given the massive reach of KPop among global youth, I wouldn't be surprised if the most well-known 안씨 as of 2025 was an "An" (a member of the group 아이브). Roh has fallen out of favor, young 노s generally go with Noh, the Rohs are usually older people. I too do long for the day where an 이 or 우 just goes with I or U, or if they must, at least Ih or Uh :)
IMO you left out the worst offender, Park. At least with 이 or 우 I can see why people would be hesitant to go the proper route, as most of the world is unfamiliar with single-phoneme names, but 박s have no excuse.
With 이, there's a pretty good alternative as well, and what's more - it's actually already in use when talking about the greatest Korean in history, Yi Sun-Shin! So much better than "Lee".
Oh there are plenty of standards, including an official one. The problem is nobody uses them. Thai writing is weird, and between the tones and the character classes and silent letters might as well just make some shit up. My birth certificate, drivers license, and work permit all had different spellings of my name on them.
IIRC, the road signs for “Henri Dunant Road” were spelled differently on either end, which was ironic, because at least that did have a canonical Latin form.
Readit News
The ō in Hepburn could correspond to おう or おお or オー. That's an ambiguity.
Where does Hepburn disambiguate?
In Japanese, an E column kana followed by I sometimes makes a long E, like in 先生 (sen + sei -> sensē). The "SEI" is one unit. But in other situations it does not, like in a compound word ending in the E kana, where the second word starts with I. For instance 酒色 (sake + iro -> sakeiro, not sakēro).
Hepburn distinguishes these; the hiragana spelling does not!
This is one of the issues that makes it very hard to read Japanese that is written with hiragana only, rather than kanji. No word breaks and not knowing whether せい is supposed to be sē or sei.
There are curiosities like karaage which is "kara" (crust) + "age" (fried thing). A lot of the time it is pronounced as karāge, because of the way RA and A come together. Other times you hear a kind of flutter in it which articulates two A's.
I have no idea which romanization to use. Flip a coin?
Slightly off-topic, but “karaage” (kara + age) isn’t “crust + frying.”
The kara comes from a country name and refers to a style of cooking — it’s a “country-name + cooking method” compound.
this is the commonly accepted explanation, though whether it’s strictly historical or a later interpretation is still debated.
If you fry something without coating it, that’s usually called “su” (plain) + “age” (frying) instead.
My understanding is that the exact etymology is unknown. It's often written with the letter that references the tang dynasty, but the thing is there's no particular reason to think the Chinese introduced the style of cooking to Japan - although it is true that there was such a thing as fried chicken in 7th century China!
Another kanji-ization of the word uses the kara from karate (meaning air or empty, in karate it's "empty hand") and I find this equally plausible as karaage is fried with a very small amount of batter ("in air").
Either way they're both essentially competing "kanji backronyms" seeking to retcon an existing word as spoken; there's no real right or wrong answer.
Still, that sort of thing in general still leaves room for it having been word play. Like tempura being originally from Portuguese, having nothing to do with 天.
Japanese spelling often plays gaslighting head games.
While it is sometimes difficult to discern the combined E and I sound, especially for non-native speakers, the word 先生 (sensei) is technically pronounced "sensei" and should be spelled that way to distinguish it from words with long E sounds, such as ええ (ee) and お姉さん (oneesan). Similarly, the OU in 東京 (toukyou) and the OO in 大きな (ookina) are different and should be spelled differently. I hope this helps.
EDIT: Added a comma.
But spelling out and singing aren't normal speech. Spelling/singing can break apart diphthongs, like NAI becomes NA-I.
生 is not written with い due to the /e:/ having a different sound from that one in from おねえさん. It does not (when you aren't spelling). It is written the way it is for ancient historic reasons.
> Similarly, the OU in 東京 (toukyou) and the OO in 大きな (ookina) are different
No, they are't.
> I hope this helps.
こう言うバカな戯言は少しも誰にも役に立つはずないんだぜ。
If you're an English speaker, you can be forgiven for a very stereotypical trait of the English accent. English speakers have a real hard time with the /e/ or /e:/ sounds as well as the /o/ and /o:/ sounds. Most English dialects don't have either a monophthong /e/ or /o/. Both the long and short tend to get heard as /eɪ/ and /oʊ/.
French enchanté /ɑ̃ ʃɑ̃ te/ is heard and borrowed as /ɑn.ʃɑn.teɪ/. German gehen /ge:n/ is heard as "gain" /geɪn/. And Japanese /o:/ and /ou/ both get heard as /oʊ/.
It's arguably a minimal pair in Japanese: 負う /ou/ (to carry), 王 /o:/ (king).
Like how many people end up with the same romanized name while being distinct in other alphabets. Then discrepancies between the different systems because they usually are sloppy on the handling of these matters.
Now that most stuff is electronic, these small differences can have wider effects and be a PITA to fix.
That said, this is far from the most important problem in Japanese pronunciation for westerners, and at speed the distinction between them can become very subtle.
You need to know previously the word to write from Hepburn to Kana when "ō" is present because data is lost in such transliteration from おう or おお or オー to Hepburn.
The internet is full of romanji written incorrectly with "o" alone when it should be "ou" or "oo" due "ō" ASCII conversion errors at one moment.
(The sooner a beginner embrace Hiragana and Katakana, the better)
Deleted Comment
To me, Hepburn’s strength relative to the old government romanization is that it increases the likelihood that an English speaker will make approximately the right sound when reading some Romaji, and that people seem to prefer it in general.
Deleted Comment
https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...
Some Japanese words entered Russian not directly, but through English. In these cases, the word is first romanized using Hepburn, and then adapted to Russian using English-to-Russian rules. A classic example is 寿司, which Polianov would render as суси (susi), but Russians mostly know as суши (sushi). Then there are words which actually do faithfully follow Polianov, as in 新宿, which is written as Синдзуку (Sindzuku) instead of Шинджуку (Shinjuku).
1. It's "Polivanov", not "Polianov".
2. It's "Синдзюку", not "Синдзуку".
Another example of JP→EN→RU is Nintendo's character Yoshi: By Polivanov, it should have become "Ёси" but since it came to RU via EN, it is written as "Йоши".
しんじゅく (Cиндзюку, Sindzyuku) is an interesting case, as it has both し and じゅ in it. This is where Polivanov is similar to Kunrei. OTOH, Fukushima is cyrillized as Фукусима (Fukusima), where the ふ is a fu in Hepburn, hu in Kunrei and fu in Polivanov but し is not shi as in Hepburn, but si as in Kunrei.
For people not interested in learning Japanese, however, a unified romanization could have its benefits. It just never struck me as particularly inconsistent to begin with, even after so many years living there.
I’ve met a few students of this textbook system when I was on exchange and my impression was that they were very skilled at Japanese for the amount of time they’ve been a student and what they told about their seniors was they pick up kanji fast, since they already know the words.
The big problem of course is that it is completely incompatible with other schools. Where do you place them when they go on exchange? With the n3 or n5 students?
Anyway, I always thought it was interesting that the exact antithesis of RTK* exists and works.
*RTK or “remembering the kanji” is a system that teaches all kanji before student learn their first word. It’s quite popular online as it lends itself very well to solo studying.
One thing I have found over the years, I have never met a foreigner living in Japan who has used it extensively. (Many were aware of it, but few used it heavily.) However, there is a lively community of online learners who use it. (Don't read that as a judgement against using it; this is simply an observation.)
I was surprised to read this part:
I have never heard this description before. I always thought it was a learning aid to use mnemonics to remember the meaning of individual kanji. If someone can complete all volumes of RTK before "learn[ing] their first word", I would be stunned. It would be a feat of super-human level of memorization and recall. That said, the Internet is a huge place with billions of people. There will be somebody, somewhere who took this path and is happy to tell you about their success using it.Like with a lot of things like this, if you learn for long enough the differences in the major approaches work themselves out.
From what I understand this isn't the first time they've made some kind of change to orthography, I remember reading something about updating offical use of certain kana to reflect more modern pronunciations. It wasn't a dramatic change.
It's interesting to see some countries just have this centralised influence over something like how their language is written as they're the main ones speaking it, as opposed to English.
One weird thing that I have noticed when Japanese native speakers write emails in English: Why don't they use basic spell check? I'm talking about stuff as basic as: "teh" -> "the". Spell checkers from the early 1990s could easily correct these issues. To be clear, I rarely have an issue to understand the meaning of their emails (as a native speaker, it is very easy to skip over minor spelling and grammar mistakes), but I wonder: Why not spell check before you send?
https://en.wikipedia.org/wiki/Compose_key
I find this often with apps and websites, and I speak/write British English (or attempt to).
Why effort is put into making a worse interface is baffling.
On mobile I just switch to the hiragana keyboard, but that obviously isn't a sane option on desktop unless I'm clicking all the characters with a mouse?
Native Cantonese speakers in Hongkong have similarly diverse input methods. I've even seen tiny digital draw pads at the public library. It is pretty exciting (to me!) to watch an elderly person furiously scribbling away on the pad, inputting traditional Chinese charaters to search something on the Internet or in the media catalog. I think it is very cool that public library makes a strong effort to empower all types of users.
It might sound complicated at first, but you can do it pretty fast once you get used to it.
https://learn.microsoft.com/en-us/globalization/input/japane...
Standard Qwerty keyboards are well supported, you'll need to either check the key shortcut to switch between inputs or do it with the mouse if it's infrequent enough.
People using it daily will tweak a lot more, have a straight to IME and straight out of IME key instead of the default switching pattern, potentially add more tweaks to always have half-width space and ponctuation whatever the mode they're in etc., but that's a rabbit-hole you'll be free to fall into.
BTW the reverse works well enough: Windows has a specific mode to force US ANSI on JIS layouts and still use the additional japanese keys. Kinda fun they felt the need to leave that mode in.
Assuming Microsoft's Japanese IME is still a dumpster fire, and the Google one has not succumbed to Googleshitification, that would be a way to go.
To enable the Microsoft IME there are some rituals to go through like adding the Japanese language and then a Japanese keyboard under that. It will download some materials, like fonts and dictionaries. A reboot is typically not required, I think, unless you make Japanese the primary language.
Once you have the keyboard, LeftShift + LeftAlt chord goes among the input methods. Ctrl + CapsLock toggles hiragana/romaji input. I think these are the same for Google or MS input.
big if true, jesus christ microsoft
Deleted Comment
(The bugs I've experienced: it doesn't properly disable itself during video games, despite claiming to do so; sometimes the popup seem to come up when I swear I didn't press the shortcut keys; rarely, the popup gets stuck on screen and needs to be Alt+F4'ed.)
> The council’s recommendation also adopts Hepburn spellings for し, じ and つ as shi, ji, and tsu, compared to the Kunrei spellings of si, zi and tu.
I could imagine si, zi and tu sound closer to the spoken sounds to Mandarin speakers.
If you look at the kana, the Japanese syllabic writing system, they have this ordering: ka ki ku ke ko, sa shi su se so, ta chi tsu te to, etc. If you follow the regularity where there should be a "ti" sound there is no "ti" sound and it happens to be pronounced "chi".
One common analysis holds that the underlying phonemes really are: ta ti tu te to. Traditional Japanese grammarians usually analyzed it this way. And they were historically pronounced that way: it has arisen out of relatively recent sound change. Somewhat like how some British speakers pronounce "Tuesday" such that it sounds much like "Chews-day" to speakers of other dialects. Affrication in a fixed context. The t phoneme triggers that kind of affrication obligatorily in Japanese, before the i vowel or y glide.
Some disagree with this as overly theoretic and based excessively on historical linguistics, and they insist that sh and f and ch are distinct phonemes in Japanese. But the Japanese writing system itself treats it as if they were not.
If you are learning Japanese it makes sense to pick a system that reflects the internal logic of kana spelling. If you want to just approximately pronounce Japanese words in English then you want something that reflects the logic of English spelling.
These two goals are always in tension. Mandarin pinyin, for example, was designed to reflect the logic of Mandarin phonology in a consistent way. It's not meant to be easily pronounceable by English speakers. It's to enable Mandarin speakers to look up words in a dictionary or for students of the language to study Mandarin. Though it has ended up used as a pronunciation guide for English speakers. And that often doesn't go well; a lot of English speakers don't know what to do with the q's and x's.
Hepburn was designed to teach non-Japanese people Japanese, therefore matching well to European (especially English) sounds was considered more important.
Suggesting Japanese romanise is a fringe position these days, much much more so than in the 1880s or the immediate aftermath of WW2, and making that kind of change is much easier when you have a population going from illiterate to literate than in a modern society, so nobody's seriously considered Nihon-shiki (or its slightly modernised descendent, Kunrei-shiki) a gateway to romanising Japanese for the Japanese for a long time now.
So this is sort of an official recognition that the primary purpose of romaji is for the benefit of foreigners.
Kunrei: ki si ti ni hi mi
Hepburn: ki shi chi ni hi mi
The politics of the issue is obviously that Hepburn is older and an American system while Nihon and Kunrei are very purposely domestic (Nihon "is much more regular than Hepburn romanization, and unlike Hepburn's system, it makes no effort to make itself easier to pronounce for English-speakers" [1]). Apparently, Hepburn was later imposed by US occupying forces in 1945.
Perhaps 80 years is long enough and suitable to effect the change officially with no loss of face.
[1] https://en.wikipedia.org/wiki/Nihon-shiki
That's the thing... to some other non-English language speakers, the existing/old romanization method actually is more accurate regarding how the letters would be pronounced to them, especially coming from languages that don't have the same e.g. [ch] or [ts] sounds as written with Hepburn.
The one technical downside I would say to this change is, 1:1 machine transliteration is no longer possible with Hepburn.
xi → hsi ji → chi ci → tz'u
Edit: actually the only one I ever had an issue with was one of my homestay's names "Ryouhei"...that Ryou sound...it's like the Y stops me from rolling the R properly, so odd.
Here's the video that got me interested in Sato www.youtube.com/watch?v=XYpFL08m5fQ&list=RDXYpFL08m5fQ&start_radio=1
[1] https://en.wikipedia.org/wiki/Hepburn_romanization#Long_cons...
つ is often seen as tu or tsu.
I have been in Japan for over a decade.
The chart halfway down this blog post lays out some of the challenges once the hanyu pinyin standard was instituted in 2009:
https://frozengarlic.wordpress.com/on-romanization/
The author concludes with this observation:
So that’s why people in Taiwan can’t spell anything consistently and why all the English-language newspapers spell the same things differently. As for me, I’m giving up on trying to remember how everyone spells their name. I know lots of people, especially Taiwan nationalists, dislike having the PRC hanyu pinyin system. I dislike imposing it upon them. However, in only three weeks, I’ve found myself spelling the same thing in multiple ways and wasting time looking up how I did it last time. Since almost no one reads my blog anyway, I’ll do it the way that’s most convenient for me.
I’ll also always provide the Chinese characters so that people who can read them know who I’m talking about.
It's unfortunate but I don't think it'll get fixed any time soon. Nobody wants to be called Mr. I, O, U, An, or No. (The most common romanization for these family names would be: Lee, Oh, Woo, Ahn, and Roh.)
No country is going to force their big multinationals to change their international name they chose back in the 50s and are now known as world-wide. Personal names aren't too chaotic either, as the choice presented when choosing a romanization is limited, people can't just make stuff up on the ground. They're off, but generally in the same ways.
> Nobody wants to be called Mr. I, O, U, An, or No.
An is pretty common - given the massive reach of KPop among global youth, I wouldn't be surprised if the most well-known 안씨 as of 2025 was an "An" (a member of the group 아이브). Roh has fallen out of favor, young 노s generally go with Noh, the Rohs are usually older people. I too do long for the day where an 이 or 우 just goes with I or U, or if they must, at least Ih or Uh :)
IMO you left out the worst offender, Park. At least with 이 or 우 I can see why people would be hesitant to go the proper route, as most of the world is unfamiliar with single-phoneme names, but 박s have no excuse.
With 이, there's a pretty good alternative as well, and what's more - it's actually already in use when talking about the greatest Korean in history, Yi Sun-Shin! So much better than "Lee".
IIRC, the road signs for “Henri Dunant Road” were spelled differently on either end, which was ironic, because at least that did have a canonical Latin form.