AniDB Definition:Romanisation

The information on this page is incomplete and may not be of much use.

If you can, please help by adding to it.

The information on this page is provided as guidelines on the use of romanised Japanese (rōmaji) in AniDB. Please be aware that this is not an exact science, there are many viable solutions to the same problem, though when submitting change requests on romanised titles users are expected to adhere to the 'house style' of the database. When there is contention over a particular issue, this page will provide both alternatives. External links to Wikipedia are provided throughout for ideas and terms that might be unfamiliar.

What romanised titles are for

• Primarily, to provide a transcription of the Japanese title that is aurally recognisable and readable by a user with little or no knowledge of the language. In using roman script, this is obviously targeted at speakers of European languages, however as this consitutes a majority of the poulation of AniDB users, this is a fair restriction.

• Secondary purposes include enabling rough pronunciation of titles, providing an alternative method of searching for a japanese title, assisting novices in reading unfamiliar words, and clarification of the reading of a particular word or phrase where it might be ambiguous.

What romanised titles aren't for

• There is no requirement to be able to reconstruct the original title from romanised form. With three distinct scripts plus roman, a wide range of homophones, and typographic intricacies such as furigana usage, this is beyond the scope of a 26 letter alphabet. In all cases the Japanese title should be presented as well, a romanised form is in no way a replacement for this.
• Further more, the romanisation need not be a lossless transliteration of Japanese spelling. Though less so than English, Japanese pronunciation deviates somewhat from the phonemic spelling. As the aim is to provide an aurally recognisable transcription, it is more important to better reflect the sound than exact spelling.
• Romanised titles do not need to provide a basis for correct Japanese collation of titles. This is a technical problem that would be better handled correctly through its own system, and would interfere with the primary purpose of the romanisation.
• Romanisations need not have an 'official' status. Though both the Japanese makers and international licencees might provide a romanised title, this is irrelevant to a transcription of the japanese title - except arguably in the case of names.

Hepburn romanisation

The Hepburn romanisation system was devised for a Japanese–English dictionary, published in 1867. Despite having no official status, variations of it are used for a vast majority of transcriptions, both inside and outside Japan. Unlike the two other main romanisation schemes, it concentrates of representing Japanese phonology rather than the underlying spelling.

Table of kana romanisation

Each mora represented in the kana spelling of a Japanese word can be transcribed into roman letters according to the table below, with a few special cases that are listed in the following sections. The hiragana is on the left, katakana is on the right.

Table adpated from wikipedia article on Hepburn. Obsolete kana are shown in red.

あ a ア	い i イ	う u ウ	え e エ	お o オ	(ya)	(yu)	(yo)
か ka カ	き ki キ	く ku ク	け ke ケ	こ ko コ	きゃ kya キャ	きゅ kyu キュ	きょ kyo キョ
さ sa サ	し shi シ	す su ス	せ se セ	そ so ソ	しゃ sha シャ	しゅ shu シュ	しょ sho ショ
た ta タ	ち chi チ	つ tsu ツ	て te テ	と to ト	ちゃ cha チャ	ちゅ chu チュ	ちょ cho チョ
な na ナ	に ni ニ	ぬ nu ヌ	ね ne ネ	の no ノ	にゃ nya ニャ	にゅ nyu ニュ	にょ nyo ニョ
は ha ハ	ひ hi ヒ	ふ fu フ	へ he ヘ	ほ ho ホ	ひゃ hya ヒャ	ひゅ hyu ヒュ	ひょ hyo ヒョ
ま ma マ	み mi ミ	む mu ム	め me メ	も mo モ	みゃ mya ミャ	みゅ myu ミュ	みょ myo ミョ
や ya ヤ		ゆ yu ユ		よ yo ヨ
ら ra ラ	り ri リ	る ru ル	れ re レ	ろ ro ロ	りゃ rya リャ	りゅ ryu リュ	りょ ryo リョ
わ wa ワ	ゐ wi ヰ		ゑ we ヱ	を wo ヲ
				ん n ン
が ga ガ	ぎ gi ギ	ぐ gu グ	げ ge ゲ	ご go ゴ	ぎゃ gya ギャ	ぎゅ gyu ギュ	ぎょ gyo ギョ
ざ za ザ	じ ji ジ	ず zu ズ	ぜ ze ゼ	ぞ zo ゾ	じゃ ja ジャ	じゅ ju ジュ	じょ jo ジョ
だ da ダ	ぢ (ji) ヂ	づ (zu) ヅ	で de デ	ど do ド	ぢゃ (ja) ヂャ	ぢゅ (ju) ヂュ	ぢょ (jo) ヂョ
ば ba バ	び bi ビ	ぶ bu ブ	べ be ベ	ぼ bo ボ	びゃ bya ビャ	びゅ byu ビュ	びょ byo ビョ
ぱ pa パ	ぴ pi ピ	ぷ pu プ	ぺ pe ペ	ぽ po ポ	ぴゃ pya ピャ	ぴゅ pyu ピュ	ぴょ pyo ピョ
Extended Katakana - These are used mainly to represent the sounds in words in other languages. Most of these are not formally standardized and some are very rarely used.
			ye イェ
	wi ウィ		we ウェ	wo ウォ
va ヷ	vi ヸ		ve ヹ	vo ヺ
va ヴァ	vi ヴィ	vu ヴ	ve ヴェ	vo ヴォ
			she シェ
			je ジェ
	ti ティ	tu トゥ	che チェ			tyu テュ
	di ディ	du ドゥ				dyu デュ
tsa ツァ	tsi ツィ		tse ツェ	tso ツォ
fa ファ	fi フィ		fe フェ	fo フォ		fyu フュ

Special cases

Hepburn also has a few extra rules to deal with particular cases, the ones below the AniDB house style adheres to.

The particle spelling rules exist to reflect modern japanese pronunciation, note there are other features that Hepburn does not attempt to reflect, for instance the frequent dropping of the vowel /u/ (です is only pronounced 'desu' by kids), largely because there's no easy rule that could always be applied. The 'small tsu' rules reflect the fact it used in two rather different ways, and the syllabic n case is to deal with the problem that transcription might be ambiguous in a few cases.

Particle は as wa

Intro to は by Tae Kim

This rule is basically accepted by everyone, generally only ignored in error.

When used as a particle, transcribe は as 'wa' rather than 'ha'

• Better represents the pronuciation
• Common practice everywhere

Particle へ as e

Intro to へ by Tae Kim

Sometimes contested, as romanisations that ignore this rule are somewhat more common. Use 'e' in preference, but if adding an anime title where 'he' is sometimes used, add that alternative as a synonym.

When used as a particle, transcribe へ as 'e' rather than 'he'

• Better represents the pronunciation
• Established Hepburn rule, and widespead usage by those who follow transcription rules strictly
• Titles will save one character per へ particle

Always transcribe へ as 'he', even when particle

• Once less rule to remember
• Common practice amoungst fansubbers
• Some titles including the particle へ are generally called by names romanised with 'e' by fans

Template:A2 also particularly resistant to using more sensible punctuation.

• There's not much difference between pronunciation of /he/ and /e/

Particle を as o

Intro to を by Tae Kim

Often contested, many romanisations in the wild will use 'wo'. Feel free to use either, but do not make change requests that just change the particle... If adding as an anime title, include a synonym with the alternative romanisation.

When used as a particle (you won't ever see it used in a normal word, so this means always (except when you do, as in names, or old words sometimes written in kana ... (like this? http://ja.wikipedia.org/wiki/ちみもりを also, use the talk page, you tards))), transcribe を as 'o' rather than 'wo'

• Better represents the pronunciation
• Established Hepburn rule, and widespead usage by those who follow transcription rules strictly

Always transcribe を as 'wo', even when particle

• Once less rule to remember
• Common practice amoungst fansubbers
• Many titles including the particle を are generally called by names romanised with 'wo' by fans

Template:A2

Template:A2

Template:A2

Discussion: 2004.06 / 2004.06 / 2005.07 (warning: profanity)

っ when geminate consonant

Really a very simple rule, complicated by one particular case. When っ is indicating a stop, the easy way to show that in the roman alphabet is with a doubled consonant. However for っち/っちゃ/っちゅ/っちょ the cluster tch is a probably a better transcription than cch (which is also confused by use in Italian) - but which is used tends to come down to individual words, which makes applying a general rule very difficult.

When part of a word, transcribe っ by doubling the following consonant, except っち as 'tchi' and similar

• May better represent pronunciation
• Some common words are best known with a 'tch' transcription

Template:A2 The toys are best known as Tamagotchi, the spelling 'tamagocchi' not used.

Template:A2 (tatchi) is a pun on たっちゃん Tat[suya]-chan, but the っちゃん ending can be used for any name.

When part of a word, always transcribe っ by doubling the following consonant

• Once less rule to remember
• Some common words are best known with a 'cch' transcription

Template:A2 Ecchi has been borrowed back into English, and almost always spelt with the 'cch' - though this particular title is arguably just 'Futari H'.

っ when exclamation

Commonly either given as an exclamation mark or just dropped, the former is preferable.

Transcribe っ at the end of a word as '!', unless followed by one anyway, in which case drop

• The っ as suprise/intonation marker is broadly equivalent to an exclamation mark
• !! is っ! is ! semanticly, typography isn't important for transcriptions

Template:A2

Template:A2

Discard っ at the end of a word in transcription

• Not exactly crucial, is it. Is it? (punctuation joke, sorry)

Transcribe っ at the end of a word as a trailing 'h'

• More appropriate for some endings than others, is context dependant, 'ah' is sensible, 'ih' is just odd
• Potential for confusion with the habit of transcribing long vowels with an h, 'oh' could be おう or おっ

ん before vowels as n'

Accepted.

Deviations from Hepburn

These are rules in Hepburn that the AniDB house style does not follow, for reasons given.

Macron usage for long vowels

Accepted.

ん before labial consonants as m

Accepted.

Other orthography issues

Spacing

Separate each 単語 and 助詞.

Separate 隊 except when part of another word (e.g. 軍隊). Don't hyphenate.

A good way to check if it is part of another word or not, is to use a Japanese-English translator such as Babel Fish to do a translation to English. Japanese words meant to be together will often produce a single English word while the Japanese words meant to be separate will often produce separate English words.

Honorifics

(temp rule not final) Honorifics that are no standalone words are added with '-' Otherwise just seperate them by spacing

ie.:
AA! Megami-sama
but
Arete Hime

たち

Always split from the associated word and set '-'

ie.:
Elf wo Karu Mono-tachi

Capitalisation

Use an initial capital letter for 単語.