Japanese transliteration: ja-Latn-hepburn

Frank Bennett biercenator at gmail.com
Sun Sep 6 02:14:49 CEST 2009

Here is a submission for ja-Latn-hepbun.  I will send submissions for
ja-Latn-kunrei and ja-Latn-nihon shortly


Request for variant registration

1. Name of requester: Frank Bennett

2. E-mail address of requester: bennett at law.nagoya-u.ac.jp

3. Record Requested

    Type: variant
    Subtag: hepburn
    Description: Hepburn romanization.
    Prefix: ja-Latn

4. Intended meaning of the subtag:

    Indicates that the target content is Japanese text, romanized using a
    method derived from that first devised by the Society for the
    Romanization of the Japanese Alphabet in 1885, and popularized
    through the publication of a Japanese dictionary by J.C. Hepburn
    in 1886.

    The common characteristic of Hepburn romanization in its many
    variants, apart from the name, is an emphasis on approximating
    Japanese _pronunciation_ using English or European spelling
    conventions.  Hepburn romanization does not attempt to parallel
    or transcribe the Japanese logographic scripts (hiragana or katakana).

5. Reference to published description of the language (book or article):

         J.C.Hepburn, A Japanese-English and English-Japanese
Dictionary, 3rd ed., 1886.

         Revised Hepburn: ALA-LC Romanization Tables (available for download)


6. Any other relevant information:

    One of the reasons for the large variety and lack of discipline in
    Japanese romanization schemes is the simplicity of Japanese
    phonetics.  For a given Japanese word, there will be several more
    or less obvious ways of transliterating it into latin characters.
    All such schemes lose such a large amount of information when
    compared with the original text that it is difficult to make a
    persuasive argument that one scheme is significantly better than

    The problem of information loss is particularly severe in the case
    of Japanese.  Whereas in Chinese, the Han characters each have
    particular, fixed pronunciations, in Japanese these often have
    multiple readings.  This, together with a limited syllabary,
    results in a crowded namespace with many homonyms.  The result is
    an emphasis on visual form in much discourse; people in
    conversation can often be heard to describe the Han characters of
    particular words to one another for clarity (i.e.  "kome-hen no
    seikou", meaning "the word pronounced 'seikou' that starts with a
    character containing the 'rice' radical").

    The extremely loose connection between the roman transliterated
    form of a text and its original form has meant that romanized
    script is used only for very short phrases, where the intended
    meaning is often clear from the context, or in combination with
    a translation (as in many academic citation systems), where the
    translation provides a hint of the meaning of the transliterated
    phrase.  In both cases, variances in the transliteration do not
    seriously impede readability, and therefore, both by intention
    and by accident, they have proliferated.

    By the same token, for many tagging purposes, identifying text as
    "Hepburn romanization" will be sufficient, and more precise
    description would be counter-productive (because most members of
    the population are indifferent to the small differences between
    the variants).  If for particular purposes a need arises to tag
    specific, well-defined subvariants of Hepburn, they can be added
    in future.

More information about the Ietf-languages mailing list