Rōmaji or Rōmazi

by Edwin Oldfather Reischauer

[Reprinted from Journal of the American Oriental Society, Volume 60, Number 1, Pages 82-89.]

 

Rōmaji or Rōmazi

Edwin O. Reischauer
Harvard University

In the pages of this journal, Mr. Carr has brought up the vexatious problem of Japanese Romanization (Rōmaji), arguing that “there is every reason, scientific and practical, for going over to the Kokutei” Romanization and urging that “discussion regarding its adoption be opened at the first opportunity.”^[1]

In 1885, a group of interested Japanese scholars, with the aid of Hepburn, a distinguished American missionary, and other foreigners, devised a system of Romanization for the Japanese language, which was adopted and still is used by the Nihon Rōmaji Hirome Kai (Japanese Society for the Propagation of Rōmaji). In time this system came to be known by Hepburn’s name, although it would be more fitting to call it simply the Old Romanization.^[1a] In 1886 a movement was started by some natural scientists and other Japanese scholars for the reform of this system. The result was Nipponshiki (Japanese style) Rōmaji, or, according to the Romanization of this system, Nipponsiki Rōmazi. In 1937, a cabinet decree adopted for government use a system called Kokutei (Official) Rōmazi, which conforms to Nipponsiki in most of the important points of dispute between it and the earlier system.

The following chart of the differences between the three systems of Romanization will help to clarify the problem for those not familiar with the various Romanization systems or with Japanese phonetics.

Kana	Phonetic Transcription	Old Romanization	Kokutei	Nipponsiki
クヮ	ka	ka	ka	kwa
シ	ʃi	shi	si	si
シャ	ʃa	sha	sya	sya
シュ	ʃu	shu	syu	syu
ショ	ʃo	sho	syo	syo
チ	tʃi	chi	ti	ti
ツ	tsu	tsu	tu	tu
チャ	tʃa	cha	tya	tya
チュ	tʃu	chu	tyu	tyu
チョ	tʃo	cho	tyo	tyo
フ	Fu	fu	hu	hu
ジ	dȝi	ji	zi	zi
ジャ	dȝa	ja	zya	zya
ジュ	dȝu	ju	zyu	zyu
ジョ	dȝo	jo	zyo	zyo
ヂ	dȝi	ji	zi	di
ヅ	zu	zu	zu	du
ヂャ	dȝa	ja	zya	dya
ヂュ	dȝu	ju	zyu	dyu
ヂョ	dȝo	jo	zyo	dyo
ンバ	mb…	mb…	nb…	nb…
ンパ	mp…	mp…	np…	np…
ンマ	mm…	mm…	nm…	nm…

The fundamental difference between the systems is that the Old Romanization is a simple broad transcription based on the normal value of the vowels as in Italian and the consonants as in English, whereas the other two Romanizations are not phonetic transcriptions but are rather phonemic Latin orthographies which take the phonemic structure of Japanese rather than Occidental concepts of letters values as a starting point. In other words, the Old Romanization is a simple Latin transliteration designed for the transcription of single Japanese names and words when writing in a foreign language, while the other two systems are really spelling systems designed for use by Japanese wishing to write the Japanese language in the Latin alphabet. Consequently, the Old Romanization is obviously designed primarily for use by foreigners and the other two systems for use only by those who know the Japanese language.^[2]

The nature of this difference may be illustrated by the T series of Japanese syllables, which are ta tʃi tsu te to. Since tʃi and tsu take the place of ti and tu in the Japanese phonemic system, they can naturally be Romanized as ti and tu in a system of Japanese orthography with no danger of confusion between the syllables they represent and the syllables normally represented by ti and tu in most other languages. However, in a transcription system designed for foreigners, especially for those unfamiliar with the phonetic structure of Japanese, these two syllables cannot be Romanized as ti and tu without causing serious confusion. Therefore, the Old Romanization, in accordance with its own principles, naturally represents them as chi and tsu.

Whatever may be the objectives of the many groups supporting the various Romanization systems, there can be no doubt that these systems differ radically in character and are designed, whether consciously or not, for two very different purposes. On the one hand, the Old Romanization has proved to be a very practical transcription. Although it may seem to rest on rather arbitrary standards of letter values, it has the great weight of decades of almost universal acceptance on its side, and the simple rule, “the vowels as in Italian, the consonants as in English,” has been singularly successful in a geographic area where English is the undisputed international language. On the other hand, Kokutei Rōmazi is vastly superior to the Old Romanization as a Japanese Latin orthography. However, for this purpose, Nipponsiki is better than Kokutei in that it preserves a few important differences in kana spelling lost in Kokutei.^[3]

If the Old Romanization is a good broad phonetic transcription and the other two systems are good phonemic orthographies, the question naturally arises as to which is really wanted, a broad phonetic transcription for foreigners or an orthography for Japanese. From the point of view of foreigners, there can be no doubt but that the former is desired. Mr. Carr states that Kokutei has practical advantages for the student of Japanese, but there are grounds for serious doubt of this assertion,^[4] and, in any case, the argument is quite beside the point. For each student of the Japanese language there are hundreds or thousands who read and write about Japan, and it is the latter, rather than the former, who need a system of Romanization. The argument that, no matter what the Romanization, the foreigner will not pronounce Japanese with absolute correctness, and therefore the foreigner’s viewpoint need not be considered, has been brought forward, but it is mere sophistry. The fact remains that an intelligent but uninformed foreigner confronted with the Old Romanization has a much greater chance of arriving at a comprehensible approximation of the Japanese pronunciation than when confronted with Nipponsiki or Kokutei, which for him will always be full of startling and confusing discrepancies between the spelling and the correct pronunciation. The vast majority of interested foreigners, whether they be merely members of the newspaper reading public or serious students of the Far East, desire and need a broad phonetic transcription which can be used with fair accuracy by anyone, and not a phonetic orthography, which is satisfactory only for those who know the language.^[5]

The Japanese attitude towards the Romanization of their own language is somewhat confused by a divergence of objectives and by some misconceptions. Some of the most enthusiastic proponents of each of the Romanization systems have been motivated by a desire to substitute the Latin alphabet for the present means of writing Japanese in Japan itself. For this purpose they naturally need a phonemic orthography like Nipponsiki. But this group comprises only a very small number of extremists. The Japanese people and the Japanese government do not in the least contemplate such a change, and at present there is not the slightest possibility that the Latin alphabet will displace Chinese ideographs and kana in Japan. The Japanese unquestionably want a Romanization system primarily for the transcription of individual names and words in connection with foreign relations and propaganda work among foreigners. With the exception of bilingual dictionaries, Rōmaji is used for no necessary purpose for the Japanese people. Therefore, from their point of view, as well as from that of foreigners, Rōmaji is essentially for foreign use.

Although most Japanese probably realize this fact, many have allowed themselves to be misled by Nipponsiki enthusiasts. It has been argued that the Old Romanization is not scientific, while Nipponsiki and Kokutei are. It is true that, as phonemic orthographies, they are scientific, but, as broad phonetic transcriptions suitable for use by foreigners, they are neither accurate nor scientific, and for purely phonetic purposes all three systems are completely unsatisfactory.

To the Japanese, Nipponsiki and Kokutei naturally seem both simple and practical,^[6] and, forgetting that their own phonetic background is very different from that of foreigners, they believe these systems are equally simple and practical for others. Unfortunately, to the average foreigner, who naturally does not know that in the Japanese phonetic system tʃi and tsu take the places of ti and tu, both systems are extremely difficult and confusing, whereas the Old Romanization has been proved by decades of use to be both simple and practical.

Another misunderstanding on the part of many Japanese is occasioned by the names of the rival systems. The Old Romanization has been cursed by bearing the name of a foreigner, while Nipponsiki and Kokutei have names which naturally appeal to Japanese pride. But the usual assumption that the first is foreign made and the others Japanese made is false, and, even were it true, it should not influence the judgment of Japanese scholars, especially when the use of Kokutei or Nipponsiki unquestionably militates against the primary function of any Japanese Romanization system, which is to act as a tool in spreading information about Japan among foreigners.

One very natural solution of the conflict between the two types of Romanization might be to adopt a compromise system. Kokutei itself in certain regards is such a compromise. However, when the systems are based on different principles and have different objectives, there can be no satisfactory compromise. In fact, compromise, as in the case of Kokutei, is worse than either extreme. Rōmaji must be either a phonetic transcription or else a phonemic orthography but not a combination of the two.

Some may argue that the action of the Japanese cabinet has already decided the question, but this is a serious misinterpretation of the facts. Only last February the Minister of Education admitted that a final decision had not been reached.^[7] Moreover, many if not most of the Japanese groups and institutions which are most deeply concerned in the problem and print most of the material using Rōmaji are resolutely continuing to use the Old Romanization. Furthermore, the Old Romanization is still accepted everywhere outside of Japan, and the weight of the forty or fifty years during which it has been the acknowledged standard system has given it a lead in world use over Nipponsiki and Kokutei that could not be overcome in anything short of several decades. It will be no simple task to substitute such startling Romanization as Tyōsyū for Chōshū, Huzi for Fuji, Tusima for Tsushima and zyūzitu for jūjitsu. The half-hearted support of Kokutei by certain groups in Japan will accomplish little in the face of the evident superiority of the Old Romanization, its decades of unquestioned supremacy, and the determined support it is being given by many Japanese as well as most foreigners. Kokutei has by no means won the battle. It still is decidedly the weaker opponent of the two, and any deflection from the Old Romanization on the part of this society or any other group, whether in Japan or abroad, will not tend towards uniformity, as Mr. Carr asserts, but will only lead towards greater confusion.^[8]

In conclusion, one may summarize the case for the Old Romanization as follows. Of the three systems it alone is a phonetic transcription suitable for foreigners. With the exception of a few extremists, foreigners and Japanese alike desire such a phonetic transcription and not a phonemic orthography designed to displace Chinese characters and kana in Japan itself, but many Japanese have been misled into supporting Nipponsiki or Kokutei because of misunderstandings as to the relative merits of the systems and the misconception that the Old Romanization was foreign in origin. Despite the action of the Japanese government, there is strong opposition to the use of Kokutei in Japan as well as abroad, and the Old Romanization still remains by far the most generally accepted system. Therefore, if serious confusion is to be avoided in the transcription of Japanese names and words, a strong stand should be taken by Japanese and foreigners alike against the use of Kokutei or Nipponsiki. If this is done, there is good reason to believe that before long the Old Romanization may again be officially accepted in Japan, as it is elsewhere.

---

1. JAOS 59 (1939). 99-102.

   ^[1a]  The committee which evolved this system was composed of four Japanese, Toyama Masakazu 外山正一, Terao Hisashi 寺尾壽, Kanda Naibu 神田乃武, and Yatabe Ryōkichi 矢田部良吉, and of two foreigners, Mr. Chamberlain and Mr. Eby. Dr. Hepburn and a Herr Techow were merely invited to assist the group. See A Short Statement of the Aim and Method of the Rōmaji Kai, Tōkyō 1885, p. 4-6.

2. Palmer in his Principles of Romanization (Tōkyō, 1931), p. 123, has already clearly stated this difference as follows:

   “The Hepburn system is by its very nature nothing other than a transliteration whereas the Nihonshiki (i. e., Nipponsiki) is by its very nature a national orthography; in short, each system fulfils the purpose for which it was primarily intended.”

3. Another change for the worse in Kokutei is the substitution of the hyphen for ’ between the final n of one syllable and an initial vowel in the next syllable within a single word. The hyphen is best reserved for use in words which should be hyphenated on other than phonetic grounds.
4. To the student of the language the fact that Kokutei preserves the stem of the verb matsu (matu) in all forms unquestionably is gratifying, but that is of little significance, for it is generally accepted that students of the language should use kana and Chinese ideographs rather than Rōmaji in language study.
5. No person with whom I have spoken about the subject in the United States has been in favor of the adoption of Kokutei, and the only argument put forward in support of it is that, if the Japanese all use it, we may be forced to do so also for the sake of uniformity. The instructors of Japanese in eastern universities and others closely concerned with the problem of Japanese Romanization, who are listed below, all recently signed their names to a simple statement that they were in opposition to the adoption of Kokutei.

   William R. B. Acker	Freer Gallery of Art
   Hugh Borton	Columbia University
   Knight Biggerstaff	Cornell University
   Serge Elisséeff	Harvard University
   John K. Fairbank	Harvard University
   A. W. Hummel	Library of Congress
   Shio Sakanishi	Library of Congress
   Osamu Shimizu	Columbia University
   A. C. Soper	Princeton University
   Nancy Lee Swann	Gest Library
   R. Tsunoda	Columbia University
   James R. Ware	Harvard University
   Joseph K. Yamagiwa	University of Michigan

6. An often overlooked but important practical drawback to these systems from the strictly Japanese point of view is that their use undoubtedly encourages the Japanese student in his favorite mispronunciations of foreign words, which render so much of the English and other foreign languages spoken in Japan quite unintelligible to those not accustomed to Japanese mispronunciations.
7. In proof of this assertion, I offer the following statements, made in the Japanese diet on February 9, 1939 (as reported in The Japan Times Weekly, Feb. 16, 1939).

   “Mr. Kasai: … For 55 years … Romaji has been used in Japan and abroad without any difficulties.

   “But unfortunately, in recent years, Dr. Tanakadate has been trying his best to propagate a new system called ‘Romazi’ and has been doing his utmost to force it upon our people. It is indeed a crime that such an inadequate and inferior system should be forced upon our people and foreigners, when the existing system of Romaji has been used with satisfaction for over a half century. … I sincerely hope that the Minister of Education will use his discretion to put a stop on (sic) the use of this incomplete and imperfect ‘Romazi.’

   “Minister of Education, General Araki: The question of the use of Romaji has been discussed often. … I am hoping that we may be able to establish an institution for studying sound, and before arriving at a final decision, we must make careful studies of our own language and sounds.”

8. The inevitable confusion which would result can be illustrated by two well known examples. The steamship company operating the liner Chichibu Maru was induced to change the spelling of the name to Titibu Maru in conformity with Kokutei. The resulting confusion in pronunciation among the foreign passengers of the ship made the company desire to restore the old spelling, but the supporters of Kokutei vigorously opposed this. Eventually the company, in desperation, changed the name of the liner to Kamakura Maru, a name Romanized alike in all systems. The second example is afforded by a famous Tōkyō grocery store of Meidiya, which in the old Romanization would be Meijiya. Several decades ago the company began using Nipponsiki for its name, and in all these years foreigners in Tōkyō, including those who speak the Japanese language, have persisted in pronouncing the second syllable of the name as di, although this syllable does not exist in the Japanese phonetic system.