There are at least four different ways of romanizing Japanese words (that is, writing them using the Roman alphabet instead of the Japanese alphabets), and the fact that different textbooks and dictionaries use different methods of romanization, and sometimes even invent their own systems, causes some confusion as to how Japanese is actually pronounced. The Japanese themselves tend to mix different systems, even with words in the same sentence. This can be very confusing.

There seem to be two different types of romanization systems: those that try to approximate the pronunciation and those that try to imitate the Japanese phonetic syllabaries (kana) for consistency and ease of conjugation.

The systems most frequently encountered are JSL, Shin-kunrei-shiki, Hepburn, and Nippon-shiki. The system most familiar to Westerners is probably the Hepburn system, a proununciation-based system. To get an idea of the differences between the four systems, here are the various romanizations of the word for "romanization":

JSL roomazi
Shin-kunrei-shiki rômazi
Hepburn rõmaji
(note: that tilde is supposed to be a straight line, but I don't think HTML can produce that character)
Nippon-shiki rõmadi

Like I said, the most common systems are JSL and Hepburn, which are unfortunately the two with the least in common. For example:

Ohayõ gozaimasu. Watashi wa Jîbasu to moshimasu, soshite nõdo o kakitai desu yo. (Hepburn)
Ohayoo gozaimasu. Watasi wa Jiibasu to mosimasu, sosite noodo o kakitai desu yo.(JSL)
Good morning. I'm Jeeves, and I like to write nodes!

Because JSL emphasizes consistency and corresponds directly to the native Japanese phonology, it tends to ignore the actual sound of a phoneme in favor of a consistent way of representing it:

  • "si" is pronounced "shi"
  • "zyoo" is pronounced "jõ"
  • "syoo" is pronounced "shõ";
  • "tu" is prounced "tsu"
...and so forth. Those odd representations of simple sounds make it slightly easier to manipulate the words down the line, but require the student to learn how to read JSL itself.

An advantage of JSL is that it doesn't require special symbols -- long vowels are simply written as a double vowel, like ã --> aa. The major disadvantage is that the average foreigner has no idea how exactly to prounounce a word like "zyugyoo", which would be rendered "jugyõ" (class) using Hepburn.

When I write about Japanese for fellow Westerners to read, I usually use Hepburn, but with the vowel system of JSL to avoid special symbols. So if you see "deshoo", remember that the double "o" is pronounced "oh" rather than "ooh".

Non-vowel differences between JSL and Hepburn are summarized below:

JSL	Hepburn
sya	sha
si	shi
syu	shu
syo	sho
zya	ja
zi	ji
zyu	ju
zyo	jo
tya	cha
ti	chi
tyu	chu
tyo	cho
tu	tsu
hu	fu

Note, however, that the Japanese syllable "hu" or "fu" is actually pronounced somewhere between the two, like blowing out a candle or something. It's tricky.

To get around the problems of romanization, I advise simply learning the syllabaries (katakana and hiragana) and writing all your Japanese in its native script. This makes learning the language easier and causes no pronunciation or rendering confusion. Unfortuately, I don't believe that Japanese characters are supported by E2, so romanziation is our only option here.

ローマ字綴り法 - romanization

A Japanese romanization is a method of writing down Japanese in a Latin-derived alphabet system. It is commonly referred to by its slang name ローマ字 (Rōmaji), sometimes even on government-owned webpages that end in go.jp. Rōmaji is technically the Latin alphabet plus w and j by formal definition. The earliest romanizations by Dutch and Portuguese missionaries later helped linguists re-create Japanese pronunciation of the past Edo era.

Here is a brief introduction of the major Japanese romanization methods:
  • ヘボン式ローマ字 (Hebon-shiki romaji) - Hepburn-style Romanization,
    a.k.a. 標準式ローマ字 (Hyojun-shiki romaji) - Standard-style Romanization
    A romanization system based on the English representation of sounds. This romanization method is used for peoples' names in Japanese passports, and for signs on Japanese streets, according to international agreements. This system offers the least amount of confusion to people comfortable with the English writing system.

    The downside of this system is that English speakers benefit at the cost of confusion and inconvenience to speakers of all other languages. Different Western countries would prefer different romanizations. For example, when the Japanese fruit kaki was imported to Italy, they were written as "cachi" since they had a different way of transcribing sounds than the English speakers. Not only that, as apparent in the English-based romanization "sa" and "shi," sounds that are considered the same to the native speaker of Japanese are confusingly represented differently because the speaker of a Western language hears them as different consonants.
  • 訓令式ローマ字 (kunrei-siki rômazi) - Kunrei-style Romanization
    a.k.a. ISO 3602
    An official standard adopted by the Japanese government on December 9, 1954 for use in education, businesses, and pretty much anything else except where Hepburn is used. It replaced an older Kunrei Romanization which was adopted on September 21, 1937. The kunrei-siki is taught in schools, and often found in Japanese URLs.

    A detriment of using this system is the loss of one-to-one correspondance with the kana system. The letters じ, ず, ぢ, and づ are written as zi, zu, zi, zu respectively, as it is pronounced in standard Japanese dialect. However, historically and in some regional dialects, the four letters have sounds distinct from each other. As the chart below shows, the Japan-style romanization preserves this detail. Kunrei-siki can be seen as a compromise between Nihon-siki and Hebon-shiki.
  • 日本式ローマ字 (Nihon-siki rōmazi) - Japan-style Romanization
    Introduced to the world in 1885 by a physicist named Dr. TANAKADATE Aikitsu (田中舘愛橘博士), later to become an international standard. This is a very linguistically correct approach where phonemes are uniquely represented. The romanization scheme is based on the structure of the fifty sounds chart, making it easier for Japanese people to learn. It is the only system that unambiguously represents the original Japanese kana. A romanization system by the people, and for the people of Japan.

    Unfortunately, this romanization doesn't correspond well with the sound representation of any single Western language, making it difficult for all foreigners to get used to.

    Note: 日本 can be pronounced as either "Nihon" or "Nippon."

Hiragana		Hepburn		   Kunrei chart I		Japan
(Seion)
あいうえお	  a   i   u   e   o	  a   i   u   e   o	  a   i   u   e   o
かきくけこ	 ka  ki  ku  ke  ko	 ka  ki  ku  ke  ko	 ka  ki  ku  ke  ko
さしすせそ	 sa shi  su  se  so	 sa  si  su  se  so	 sa  si  su  se  so
たちつてと	 ta chi tsu  te  to	 ta  ti  tu  te  to	 ta  ti  tu  te  to
なにぬねの	 na  ni  nu  ne  no	 na  ni  nu  ne  no	 na  ni  nu  ne  no
はひふへほ	 ha  hi  fu  he  ho	 ha  hi  hu  he  ho	 ha  hi  hu  he  ho
まみむめも	 ma  mi  mu  me  mo	 ma  mi  mu  me  mo	 ma  mi  mu  me  mo
や ゆ よ	 ya      yu      yo	 ya      yu      yo	 ya      yu      yo
らりるれろ	 ra  ri  ru  re  ro	 ra  ri  ru  re  ro	 ra  ri  ru  re  ro
わゐ ゑを	 wa   i       e   o	 wa   i       e   o	 wa  wi      we  wo
  ん  		  n			  n			  n
(Dakuon)
がぎぐげご	 ga  gi  gu  ge  go	 ga  gi  gu  ge  go	 ga  gi  gu  ge  go
さじずぜぞ	 za  ji  zu  ze  zo	 za  zi  zu  ze  zo	 za  zi  zu  ze  zo
だぢづでど	 da  ji  zu  de  do	 da  zi  zu  de  do	 da  di  du  de  do
ばびぶべぼ	 ba  bi  bu  be  bo	 ba  bi  bu  be  bo	 ba  bi  bu  be  bo
ぱぴぷぺぽ	 pa  pi  pu  pe  po	 pa  pi  pu  pe  po	 pa  pi  pu  pe  po
(Renji)
きゃきゅきょ	kya     kyu     kyo	kya     kyu     kyo	kya     kyu     kyo
しゃしゅしょ	sha     shu     sho	sya     syu     syo	sya     syu     syo
ちゃちゅちょ	cha     chu     cho	tya     tyu     tyo	tya     tyu     tyo
にゃにゅにょ	nya     nyu     nyo	nya     nyu     nyo	nya     nyu     nyo
ひゃひゅひょ	hya     hyu     hyo	hya     hyu     hyo	hya     hyu     hyo
みゃみゅみょ	mya     myu     myo	mya     myu     myo	mya     myu     myo
りゃりゅりょ	rya     ryu     ryo	rya     ryu     ryo	rya     ryu     ryo

ぎゃぎゅぎょ	gya     gyu     gyo	gya     gyu     gyo	gya     gyu     gyo
じゃじゅじょ	 ja      ju      jo	zya     zyu     zyo	zya     zyu     zyo
ぢゃぢゅぢょ	 ja      ju      jo	zya     zyu     zyo	dya     dyu     dyo
びゃびゅびょ	bya     byu     byo	bya     byu     byo	bya     byu     byo
ぴゃぴゅぴょ	pya     pyu     pyo	pya     pyu     pyo	pya     pyu     pyo
くゎ    				 ka			kwa
ぐゎ    				 ga			gwa

(Special sounds)
Kunrei chart II
    sha shi shu sho
            tsu
    cha chi chu cho
             fu
     ja  ji  ju  jo
         di  du
    dya     dyu dyo
    kwa     gwa
     wo

Hepburn special sounds
ファ フィ フ  フェ フォ 	fa fi fu fe fo
ヴァ ヴィ ヴ  ヴェ ヴォ 	va vi vu ve vo
タ  ティ テゥ テ  ト  	ta ti tu te to
ダ  ディ デゥ デ  ド  	da di du de do
テュ デュ          	tyu dyu
* Kunrei chart II (above) is used only for international reasons or other circumstances where using chart I is undesireable.


Various Romanization Schemes - Details

In the bulleted list below, the separation section indicates a case where a syllabic N in Japanese is proceeded by a vowel or a semivowel, which is different from having the two sounds coming from a single syllabary letter, both conceptually and phonetically. To indicate this separation, there may be a hyphen or an apostrophe inserted inbetween. The syllabic N section tells whether a syllabic N is always written as an n, or as an m when proceeded by a b, m, or p for the romanization. The long vowels section tells us how to represent a long vowel in the ordinary case. A long vowel may be a result of two vowels from different morphemes being next to each other, in which case the pronounced long vowel is represented by two separate letters, e.g. the word for "yellow": きいろ → kiiro. The geminate consonant section tells us how to represent a consonant that is pronounced over a longer-than-usual duration of time. At the end of each list below, source(s) and reference(s) are appended.
  • Hepburn romanization for road signs
    Underneath a Japanese place name on a road sign are Hepburn romanizations that are at one-half the height of Japanese kanji-kana writing. English words are used for common words like "street" and "bridge."
    • separation: hyphen.
    • syllabic N: n.
    • long vowels: unmarked.
    • geminate consonant: initial letter doubled, except for long ch which is tch.
    • http://www.kictec.co.jp/inpaku/iken%20keikai/syasin/hebon/romaji.htm
  • Hepburn romanization for passports
    Japanese passport names are required to be in Hepburn romanization, with two exceptions: (a.) when a long o is in your name, it could be represented as oh (since April 1, 2000), or (b.) when your name is of non-Japanese origin and a different romanization is more suitable, e.g. フランシス → "Francis" as opposed to "Furanshisu."
    • separation: apostrophe, hyphen, or none.
    • syllabic N: n / m.
    • long vowels: unmarked, except long o can be either o or oh.
    • geminate consonant: initial letter doubled, except for long ch which is tch.
    • http://www.boston.us.emb-Japan.go.jp/passport.html
  • Japan-style romanization
    • separation: apostrophe.
    • syllabic N: n.
    • long vowels: Macron is used, i.e. ā, ī, ū, ē, ō.
    • geminate consonant: Initial consonant repeated.
    • http://xembho.tripod.com/
  • New Kunrei romanization
    • separation: apostrophe.
    • syllabic N: n.
    • long vowels: circumflex is used, i.e. â, î, û, ê, ô, with an exception: if the long vowel is capitalized, then it may be represented with two letters, i.e. Aa, Ii, Uu, Ee, Oo.
    • geminate consonant: Initial consonant repeated.
    • http://xembho.tripod.com/
  • Old Kunrei romanization
    • separation: hyphen.
    • syllabic N: n.
    • long vowels: Macron is used, i.e. ā, ī, ū, ē, ō.
    • geminate consonant: Initial consonant repeated.
    • http://xembho.tripod.com/
  • JSL romanization
    This romanization method comes from the textbook series Japanese: The Spoken Language by Eleanor H. Jorden (Yale University Press, 1987).

    This method describes everything from the Japanese accent to the nga ngi ngu nge ngo sounds - a level of detail not expressible in basic kana writing systems.* It expresses consonants much like how Kunrei-siki does. Unusually, it represents long vowels with two letters as opposed to representing it as a letter and a diacritical mark. Usage of diacritical marks are reserved for describing accents.

    The downside of this romanization is that it's not popular, not intuitive, and you might as well learn Japanese kana instead.

    * Japanese linguists have developed kana-based ways to write as much phonetic detail as JSL does, but it isn't supported in Unicode.

Character Entitties

Here's a bunch of HTML character entities for romaji use in writeups. Note that the characters with circumflex ( ˆ ) can be directly copy-and-pasted into a writeup without using the entity form because they are Latin-1 characters. The characters with macrons ( ¯ ) should be in entity form so it is more likely to be displayed correctly in other people's browsers. See also E2 Japanese conventions for the appropriate romanization of Japanese node titles.
Circumflex
 = Â	â = â
Î = Î	î = î
Û = Û	û = û
Ê = Ê	ê = ê
Ô = Ô	ô = ô

Macron
Ā = Ā	ā = ā
Ī = Ī	ī = ī
Ū = Ū	ū = ū
Ē = Ē	ē = ē
Ō = Ō	ō = ō

Log in or registerto write something here or to contact authors.