I fell into a Wikipedia rabbit hole about romanization standards. And in the process, I’ve realized that the way I write my wiki is… not exactly the norm.
I think most students of Japanese language will be most familiar with the Hepburn romanization. In name, if not in practice. Because why would you bother learning a language in the wrong writing system… But the name ‘Hepburn’ I think is widely recognized.
The Hepburn standard has changed over time. Most notably, in how ん (’n’) is romanized. Copying Wikipedia’s explanation (because IANAL–I am not a linguist): “written as n before consonants, but as m before labial consonants: b, m, and p”. This gave rise to 先輩 (せんぱい) being romanized as ‘sempai’ rather than ‘senpai’. I can’t really explain why this feels wrong, but the standard was effectively changed over time to favor the latter.
So the Hepburn standard is flexible and adapts to changing perceptions, great, why do I not apparently stick to it?
There are still parts of Hepburn that feel wrong. For one, as an American, I am naturally predisposed to distrust accent marks. Apparently the line over vowels is called a macron, which I surely will not remember by next week. Hepburn calls for long vowels to be romanized using macrons. For example: 勉強 (べんきょう) as ‘benkyō’.
Actually it’s pretty complicated, and Wikipedia makes it seem like it’s about the “morpheme boundary”. Which seems to mean kanji; if the long vowel is built into the singular kanji then it should be expressed with a macron. Otherwise as duplicated letters. But Wikipedia makes an example of 湖 (みずうみ) just a few paragraphs later, and suggests the romanization of ‘mizuumi’ rather than ‘mizūmi’.
‘Kneejerk reaction,’ you shout, I know, but let’s just duplicate letters. It’s simpler that way. I like simple.
There’s also the cases of vowels elongated by different vowels. 計画 (けいかく) has the け (‘ke’) elongated by a い (‘i’) instead of a え (’e’). That’s the norm in Japanese. Hepburn calls for this long vowel to be romanized as ’ei’. Fair enough.
自動販売機 (じどうはんばいき) has the ど (‘do’) elongated by a う (‘u’) instead of a お (‘o’). Again, that’s the norm. Hepburn calls for this long vowel to be romanized as… ‘ō’?
By romanizing as ‘jidouhanbaiki’ instead of Hepburn’s recommended ‘jidōhanbaiki’, I have captured in shorthand notation the correct Japanese spelling.
There’s also a point to be made about the abnormal parts of Japanese, i.e. the words that violate the above norm. 遠い (とおい) is weird because it does elongate the と (’to’) with a お instead of a う. See also 王 (おお). Hepburn rolls this into the above: ‘ō’. It feels right to me for the romanization to be similarly exceptional: ‘oo’.
So far, apparently my preferences are aligned with a semi-standard called wapuro (from word processing; a system for taking user input for computer word processing). Which I am happy to stand by.
There are some other discrepencies between what I write and what the standards suggest. Mostly these are the edge cases involving ん. If followed by a vowel, as in 小泉 純一郎 (こいずみ じゅんいちろう), the romanization can become ambuguous. Does ’ni’ mean んい or に? Hepburn calls for inserting apostrophes, which I think is hideous. Wapuro also requires these apostrophes though, because users must be able to input both possibilities.
Luckily it’s a non-issue for me, because I have never seen words like this. It’s almost exclusively names featuring this problem. (Including my example above.) My standard has been:
- Introduce a person with their name as they would have spelled it. Which would obviously be in kanji for Japanese individuals. (And unless/until his son also becomes PM, Koizumi unambiguously refers to Junichirou.)
- Introduce a person with a link to a page that has space to fully and appropriately explain the name.
There are specific words that have already been adopted into English, and it’s difficult to justify reversing course for a specific romanization style. 将軍 (しょうぐん) by my standard is ‘shougun’, but people generally expect ‘shogun’. Seems worthwhile to mention that Hepburn recommends ‘shōgun’. While that spelling is unwelcome in wider use, it’s perfectly readable for that same wider audience. Perhaps a point in its favor.
And I will never convince people that 東京 (とうきょう) is ‘Toukyou’, or that 大阪 (おおさか) is ‘Oosaka’. Perhaps another point in favor of Hepburn; by giving up the discrepency of 〜おお vs. 〜おう, the standard achieves spellings like ‘Tōkyō’ and ‘Ōsaka’ that are simultaneously readable and correct.
But then, what is the point of a standard if you can only achieve correctness by lowering the bar? I would much rather just admit to the shortcomings of a standard.
I think that Tokyo, Kyoto, Osaka, etc., are simply too recognizable by their simplified romanizations, so they get an automatic pass. Shougun is the romanization I have resolved to use because while it is now less recognizable, it’s also not a very useful word in the first place. In saying so-and-so was elected jarl of Windhelm, I have conveyed that this person held political power in a (at least semi-) democratic system. The particular title and place is relatively unimportant.
And now for a long tangent.
When are titles important? They do not have unambiguous meaning. The power vested in an office waxes and wanes; is contextual; has more to do with who someone’s contemporaries were than who they were themselves. (The powers of the U.S. House Speaker have shrunk and grown over time; such powers are vested in the office but cannot be exercised if the other members of the House cannot be kept in line; there is a monumental difference between a Speaker from the same party as the President vs. an oppositional Speaker.)
All of that context can only be captured in a separate document with a larger scope. So if a title is important, it should be introduced with a link to another page. Then the link itself can be spelled however I like.
There’s an edge case where the title has meaning but that person is the only one who ever held it. But who cares how I spell ‘Genghis Khan’, it will always be a link to Temujin.
All coronation names fall into the same bucket. It’s more common than you think. Modern examples include the papacy or the Japanese imperial throne. Similarly, there are monarchs who are addressed with different numerals depending on the reference kingdom. A trivial example: James I of England a.k.a. James VI of Scotland.
And a really fun example: There are plenty of monarchs who adopted a localized name, just as Jogaila (a Lithuanian name) took the name Władysław II (a Polish name with a lineage) upon converting to Christianity and marrying Jadwiga. (Important context: their marriage established the Polish-Lithuanian Commonwealth.) But retrospective histories have decided on ‘Władysław II Jagiełło’, a surprise combination. This sequence demonstrates just how superficial and unimportant coronation names are.