>From JAT Bulletin 11 (February 1986)

Newsletter of the Japan Association of Translators

Updated March 1997

Some Notes on Rômazi

J. Marshall Unger


The system of romanization set forth in 1885 by the Rômazi Kai and named

after James Curtis Hepburn, who later adopted it in the third edition of

his Japanese-English dictionary, has, contrary to popular belief, no

official standing in Japanese law. MacArthur decreed its use on road signs

and such on 3 September 1945, most likely because he had been led to

believe that the Kunrei system (so called because it was promulgated by

Cabinet Order [kunrei] No. 3 of 21 September 1937) was tainted by prewar

nationalism. Although Cabinet Instruction No. 1 of 9 December 1954 made

Kunrei romanization the official system a second time, the National

Railways and the Gaimusyô persist in using Hepburn (they are protected from

litigation because an "instruction" is not a "law"), and have cowed some

quasi-official private enterprises, such as the JTB, into following suit.

Nonetheless, the only officially sanctioned romanization is the Kunrei

system. Even those who wish to stick with Hepburn should therefore be

familiar with its use and rationale.

Kunrei romanization is a slightly modified version of the Nippon-siki

romanization proposed in 1885 (the same year as Hepburn) by Tanakadate

Aikitu. The chart near the front of the unabridged Kenkyûsya Wa-ei Ziten

shows the differences among all three systems. The Navy and Army adopted

Nippon-siki romanization in 1927 and 1928 respectively, and it was probably

this that set MacArthur against the Kunrei system nearly two decades later.

Kunrei romanization is more appropriate for the representation of Japanese

in latin letters than Hepburn because it more accurately reflects the

morphophonemic structure of Yamato- kotoba and kango, which comprise the

two most important strata of the Japanese lexicon. For example, in Kunrei,

verbs forms such mati, matu, matô and osi, osu, osô follow the same pattern

as yomi, yomu, yomô and so on; rendaku compounds such as makizyaku <

maki+syaku and roppun < roku+hun are parallel with midarezaku < midare+saku

and roppon < roku+hon; etc. When one takes gairaigo into consideration,

Hepburn romanization appears to have the redeeming advantage of providing

an easy way to handle certain innovative morae (onsetu). Kunrei sye, tye,

zye for Hepburn she, che, je are easy enough, but Hepburn ti, tu, etc.

require special treatment in Kunrei (e.g. t'i, t'u, etc.). In many cases,

however, using the original orthography of the gairaigo instead of a

transcription of its katakana version may be preferable, since it is of a

single fixed form and there may be less than perfect consensus on the

Japanese pronunciation. In anycase, innovative morae occur infrequently.

The Kunrei spellings zya, zyu, zyo for Hepburn ja, ju, jo bother some

people because they are each need an extra letter. Again, one should keep

in mind frequency of occurrence. The syllabes that are shorter in Kunrei

(si, ti, and tu) are fair more common than those that are longer.

Some people believe that it is easier for readers to approximate the sound

of Japanese words written in Hepburn than in Kunrei romanization. This is

false. Readers who have never heard Japanese are as likely to mispronounce

a word transcribed in one system as in the other. Hepburn and Kunrei forms

are identical except for a handful of morae, and Kunrei forms such as zi,

ti, tu, and hu, though unusual by English standards, one or more look quite

natural to readers of Czech, Danish, Dutch, Hungarian, Norwegian, Polish,

Serbo-Croation, or Swedish. Conversely, forms such as cha, jo, hi, and zu

(the last two occur in both systems) are baffling or misleading to people

who can read, for example, only French, German, Italian, or Spanish. Thus,

the spelling rules of other languages are, at best, a guide; the only

rational basis for assigning Japanese allophones to latin letters is the

structure of the Japanese sound system itself. No apologies need to be

made to the English-speaking world; on the contrary, Hepburn romanization

may contribute to faulty learning of English in Japan. Japanese who study

English and have been told that Hepburn is "based on English spelling" may

assume falsely that the Japanese sounds written f-, sh-, ch- etc. are

phonetically identical to the corresponding English sounds.

Many Japanese scientists use Kunrei romanization (Tamaru Takurô's

Nippon-siki physics text was widely used for many years), but translators,

particularly in the humanities, seem to prefer Hepburn. Certainly, there

is something to be said for precedent; however, as the switchover from

Wade-Giles to pinyin for Mandarin shows, precedent is not everything.

According to Ôtuka Haruo, former head of Japan's delegation to the

International Standards Organization (personal communication), the ISO is

putting the finishing touches on the preamble to its public announcement of

a Kunrei romanization standard. (This article is largely based on the

submission of the Japanese delegation, prepared by Saeki Kôsuke and Yamada

Hisao, to ISO/TC46 SC2 at its Paris meeting of November 1977.) Technical

translators especially should take note of this development.