Romanization of Japanese

The earliest Japanese romanization system was based on Portuguese orthography. It was developed around 1548 by a Japanese Catholic named Yajiro. Jesuit priests used the system in a series of printed Catholic books so that missionaries could preach and teach their converts without learning to read Japanese orthography. The most useful of these books for the study of early modern Japanese pronunciation and early attempts at romanization was the Nippo jisho, a Japanese–Portuguese dictionary written in 1603. In general, the early Portuguese system was similar to Nihon-shiki in its treatment of vowels. Some consonants were transliterated differently: for instance, the /k/ consonant was rendered, depending on context, as either c or q, and the /ɸ/ consonant (now pronounced /h/, except before u) as f; and so Nihon no kotoba ("The language of Japan") was spelled Nifon no cotoba. The Jesuits also printed some secular books in romanized Japanese, including the first printed edition of the Japanese classic The Tale of the Heike, romanized as Feiqe no monogatari, and a collection of Aesop's Fables (romanized as Esopo no fabulas). The latter continued to be printed and read after the suppression of Christianity in Japan (Chibbett, 1977).

Following the expulsion of Christians from Japan in the late 1590s and early 17th century, rōmaji fell out of use and was used sporadically in foreign texts until the mid-19th century, when Japan opened up again.

From the mid-19th century onward, several systems were developed, culminating in the Hepburn system, named after James Curtis Hepburn who used it in the third edition of his Japanese–English dictionary, published in 1887. The Hepburn system included representation of some sounds that have since changed. For example, Lafcadio Hearn's book Kwaidan shows the older kw- pronunciation; in modern Hepburn romanization, this would be written Kaidan (lit. 'ghost tales').

In addition to the standardized systems above, there are many variations in romanization, used either for simplification, in error or confusion between different systems, or for deliberate stylistic reasons.

Notably, the various mappings that Japanese input methods use to convert keystrokes on a Roman keyboard to kana often combine features of all of the systems; when used as plain text rather than being converted, these are usually known as wāpuro rōmaji. (Wāpuro is a blend of wādo purosessā word processor.) Unlike the standard systems, wāpuro rōmaji requires no characters from outside the ASCII character set.

While there may be arguments in favour of some of these variant romanizations in specific contexts, their use, especially if mixed, leads to confusion when romanized Japanese words are indexed. Note that this confusion never occurs when inputting Japanese characters with a word processor, because input Latin letters are transliterated into Japanese kana as soon as the IME processes what character is input.

This chart shows in full the three main systems for the romanization of Japanese: Hepburn, Nihon-shiki and Kunrei-shiki:

This chart shows the significant differences among them.

Japanese is written without spaces between words, and in some cases, such as compounds, it may not be completely clear where word boundaries should lie, resulting in varying romanization styles. For example, 結婚する, meaning "to marry", and composed of the noun 結婚 (kekkon, "marriage") combined with する (suru, "to do"), is romanized as one word kekkonsuru by some authors but two words kekkon suru by others.

There is no universally accepted style of romanization for the smaller versions of the vowels and y-row kana when used outside the normal combinations (きゃ, きょ, ファ etc.), nor for the sokuon or small tsu kana っ/ッ when it is not directly followed by a consonant. Although these are usually regarded as merely phonetic marks or diacritics, they do sometimes appear on their own, such as at the end of sentences, in exclamations, or in some names. The detached sokuon, representing a final glottal stop in exclamations, is sometimes represented as an apostrophe or as t; for example, あっ! might be written as a'! or at!.

1603: Vocabvlario da Lingoa de Iapam (1603)

1604: Arte da Lingoa de Iapam (1604–1608)

1620: Arte Breve da Lingoa Iapoa (1620)

The list below shows the Japanese reading of letters, for spelling out words, or in acronyms. For example, NHK is read enu-eichi-kei (エヌ・エイチ・ケイ). These are the standard names, based on the British English letter names (so Z is from zed, not zee), but in specialized circumstances names from other languages may also be used. For example, musical keys are often referred to by the German names, so that B♭ is called bē (べー) from German B.

• Cyrillization of Japanese
• List of ISO romanizations
• Japanese writing system
• Transcription into Japanese

• (in Japanese) Hishiyama, Takehide (菱山 剛秀 Hishiyama Takehide), Topographic Department (測図部). "Romanization of Geographical Names in Japan." (地名のローマ字表記) (Archive) Geospatial Information Authority of Japan.

• Media related to Romanization of Japanese at Wikimedia Commons
• Rōmaji sōdan shitsu contains an extremely extensive and accurate collection of materials relating to rōmaji, including standards documents and HTML versions of Hepburn's original dictionaries. (in Japanese)
• The rōmaji conundrum by Andrew Horvat contains a discussion of the problems caused by the variety of confusing romanization systems in use in Japan today.