← Home

oroamij, a novel romanisation system for the Japanese language

When I was chatting to wareya today, I had an idea for a new way to romanise Japanese.

All the romanisation systems I can think of (nihon-siki, kunrei-siki, Hepburn) romanise Japanese moræ roughly the same. Each mora, from start to finish, is transliterated individually, and written left to right. If the mora is a lone vowel (e.g. あ, エ), a lone Latin vowel is used (e.g. 'a' or 'e' respectively); if it is ん (the 'syllabic n'), it is transliterated as 'n'; if it is a consonant-vowel pair, the consonant is transliterated first (e.g. the first part of し is written 's'), doubled if preceded by a sokuon, and followed by a Latin vowel (e.g. the second part of し is written 'i'); and if it is yōon digraph, the consonant of the first kana is transliterated (e.g. the first part of じゃ is written 'j'), again doubled if preceded by a sokuon, and the second part is then transliterated the same as a normal consonant-vowel pair or lone vowel.

There are some minor differences, of course. Hepburn changes the way the initial consonsant in some consonant-vowel pair kana is transliterated, such that English speakers might pronounce them more accurately (し becomes 'shi' rather than 'si'), and it also represents elongated vowels in certain places with a macron (so 東京 becomes 'tōkyō' rather than 'toukyou'), along with other minor differences. But it is not radically different from nihon-siki/kunrei-siki in how it orders the consonants and vowels in the transliteration: initial consonant first, middle consonant if any (e.g. in a yōon) second, and finally the vowel. きゅん is 'kyun' in all three systems, 日本語 is 'nippongo'.

As an experiment, this blog post proposes something more radical. Instead of writing the initial consonant first, why not write it last?

東京 is no longer 'toukyou', but 'otuoyku'; WORLD ORDER (ワールド・オーダー) is no longer 'WAARUDO OODAA', but 'AWAUROD OOADA'; and ローマ字 is no longer 'ROOMAji', but 'OROAMij' (thus the name of the romanisation system).

The word アンインストール (from the English 'uninstall') is an interesting case. In Hepburn it would be 'an'insutōru': the apostrophe is to make it clear it's アンイ rather than アニ. Transliterated using oroamij, this becomes 'a'ni'nusotour': there's a new apostrophe, because 'aninusotour' would be ナニストール!

One possible benefit of this system is that it can distinguish consonant-vowel pairs in Japanese that require two Latin alphabet consonants to make their pronunciation clear (e.g. し is 'shi' in Hepburn) from yōon (e.g. じゃ is 'zya' in nihon-siki) more clearly. We treat the English pairs like 'sh' (for し) and 'ts' (for つ) as if they were single consonants, so し is romanised 'ish', not 'ihs'. However, we do not do the same for yōon, so ぎょ is romanised 'oyg', not 'ogy'.

This system might be best considered a variant of nihon-siki, kunrei-siki or Hepburn, rather than a completely new system in its own right, given there's no specific convention for long vowels, phonetically identical but differently-written kana, similarly-written but phonetically different kana, etc.

Trying to read this system makes my eyes backtrack a bit, but I don't think it'd actually be harder for first-time learners. If this were the only romanisation system, the fact mora are 'backwards' relative to pronunciation wouldn't matter, because you'd read it mora by mora: aw-at-ish, not w-a-t-a-sh-i.

O oyim iatadik airagotuogaziamishat.