Introduction

Recently I came across this news article from the Japan Times. After a year-long deliberation, the Japanese Agency for Cultural Affairs has issued a recommendation “to replace the government’s romanization system for the Japanese language”. The current official romanisation system in official Japanese documents and media is the Kunrei style (訓令式 lit. Cabinet’s style). If ACA’s recommendation is to take place, the current Kunrei style is expected to be replaced by the Hepburn style of romanisation. The difference between these two styles may not seem much—as we will see momentarily—but the change effectively constitutes a major defeat in standardisation, universality, and elegance.

Romanisation and romanisation styles

Romanisation is the process of converting sounds, letters, or characters of a language without Latin (Roman) characters to a particular set of characters in the Latin (Roman) alphabet. Romanisation is useful since any given language can now leverage the Latin alphabet to offer a foreign reader the possibility of recognising, reading, and pronouncing a word they would otherwise not be able to.

One might argue that the International Phonetic Alphabet is better suited for this purpose–it is!—but romanisation also has the beneficial side effect of helping with inputting and processing non-Latin words in Latin-script-based systems, be them societal or computerised.

Before we dive deeper into the differences between Kunrei and Hepburn, it is necessary for us to look at one of the most basic foundations of the Japanese language and the reason why romanisation exists in the first place. The language is divided and pronounced in syllables and its native alphabet is, thus, a syllabary. As we learnt in a previous post on Japanese typography, two syllabaries representing the same sounds exist natively.

The set of sounds in the Japanese language can be expressed, broadly, as the combination of fricative, glottal, and nasal sounds (i.e. consonants) and vocal sounds (i.e. vowels), or vocal sounds alone. Tabularised, the syllabary looks as follows:

	$\phi$	k	g	s	z	t	d	n	h	b	p	m	r	y	w
$\phi$	-	-	-	-	-	-	-	ん	-	-	-	-	-	-	-
a	あ	か	が	さ	ざ	た	だ	な	は	ば	ぱ	ま	ら	や	わ
i	い	き	ぎ	し	じ	ち	ぢ	に	ひ	び	ぴ	み	り
u	う	く	ぐ	す	ず	つ	づ	ぬ	ふ	ぶ	ぷ	む	る	ゆ
e	え	け	げ	せ	ぜ	て	で	ね	へ	べ	ぺ	め	れ
o	お	こ	ご	そ	ぞ	と	ど	の	ほ	ぼ	ぽ	も	ろ	よ	を

Gojûon: Basic consonant-vowel combinations of sounds present in modern-day standard Japanese. The empty spaces in the y and w columns represent currently unused sounds. The table has been left unromanised for the sake of argument.

	k	g	s	j	t	d	n	h	b	p	m	r
ya	きゃ	ぎゃ	しゃ	じゃ	ちゃ	ぢゃ	にゃ	ひゃ	びゃ	ぴゃ	みゃ	りゃ
yu	きゅ	ぎゅ	しゅ	じゅ	ちゅ	ぢゅ	にゅ	ひゅ	びゅ	ぴゅ	みゅ	りゅ
yo	きょ	ぎょ	しょ	じょ	ちょ	ぢょ	にょ	ひょ	びょ	ぴょ	みょ	りょ

Yôon: Digraphs present in modern-day standard Japanese. They are written as the combination of the i row and the y column in the previous table. The table has been left unromanised for the sake of argument.

Kunrei style

Kunrei style romanises gojûon as the conditioned subset obtained from the Cartesian product $C \times V$ of the set of consonants $C = \{\phi, n, k, g, s, z, t, d, n, h, b, p, m, r, y, w\}$ and the set of vowels $V = \{\phi, a, i, u, e, o\}$, such that

\[G = \{(x,y) \in C \times V \cup (n,\phi) \mid y \neq \phi \}.\]

For the less mathematically inclined, this means that either every Latin consonant in the consonant set gets assigned a vowel in the vowel set or vowels stand by themselves. The following table represents the Kunrei style of romanisation.

	$\phi$	k	g	s	z	t	d	n	h	b	p	m	r	y	w
$\phi$	-	-	-	-	-	-	-	n	-	-	-	-	-	-	-
a	a	ka	ga	sa	za	ta	da	na	ha	ba	pa	ma	ra	ya	wa
i	i	ki	gi	si	zi	ti	di	ni	hi	bi	pi	mi	ri
u	u	ku	gu	su	zu	tu	du	nu	hu	bu	pu	mu	ru	yu
e	e	ke	ge	se	ze	te	de	ne	he	be	pe	me	re
o	o	ko	go	so	zo	to	do	no	ho	bo	po	mo	ro	yo	wo

Tabularised version of the Kunrei style of romanisation for gojûon.

Similarly, Kunrei style romanises yôon as the subset obtained from the Cartesian product $C’ \times P$ of the set of consonants $C’ = \{k, g, s, j, t, d, n, h, b, p, m ,r\}$ and the set of palatised sounds $P = \{ya, yu, yo\}$.

\[Y = \{(x,y) \in C' \times P\}.\]

	k	g	s	j	t	d	n	h	b	p	m	r
ya	kya	gya	sya	jya	tya	dya	nya	hya	bya	pya	mya	rya
yu	kyu	gyu	syu	jyu	tyu	dyu	nyu	hyu	byu	pyu	myu	ryu
yo	kyo	gyo	syo	jyo	tyo	dyo	nyo	hyo	byo	pyo	myo	ryo

Tabularised version of the Kunrei style of romanisation for yôon.

Both the fact that Kunrei style is strictly mathematically regular and the fact that this romanisation was pushed by the Ministry of Education in 1937 have lead the Kunrei style to be recognised by ISO in ISO 3602.

This style, like many others in other languages, has notes that cover edge cases or properties of the language. For instance, the circumflex accent in both gojûon and yôon represents long vowels in Kunrei style.

Kunrei style is a slightly modified version of the 1885 romanisation style called Nihon style (日本式, lit Japanese style), born from the necessity of Japanese culture and language to interact, do business, and compete with Latin-script-based cultures.

Kunrei style is a tool made by the Japanese language to represent Japanese sounds using the Latin alphabet. し is transcribed as si_Kunrei, which represents the sound [ɕi]. It does not represent si_English (pronounced [siː]) nor si_Spanish (pronounced [si]). Thus,

\[し = si_{Kunrei} \neq si_{English} \neq si_{Spanish}.\] \[ほ = ho_{Kunrei} \neq ho_{English} \neq ho_{Spanish}.\]

Hepburn style

Hepburn style romanises the Japanese language based on how an English-speaking person would pronounce the particular syllable. The following table represents the Hepburn style of romanisation.

	$\phi$	k	g	s	z	t	d	n	h	b	p	m	r	y	w
$\phi$	-	-	-	-	-	-	-	n	-	-	-	-	-	-	-
a	a	ka	ga	sa	za	ta	da	na	ha	ba	pa	ma	ra	ya	wa
i	i	ki	gi	shi	ji	chi	ji	ni	hi	bi	pi	mi	ri
u	u	ku	gu	su	zu	tsu	zu	nu	fu	bu	pu	mu	ru	yu
e	e	ke	ge	se	ze	te	de	ne	he	be	pe	me	re
o	o	ko	go	so	zo	to	do	no	ho	bo	po	mo	ro	yo	wo

Tabularised version of the Hepburn style of romanisation for gojūon. Marked in red are changes compared to Kunrei style.

	k	g	s	j	t	d	n	h	b	p	m	r
ya	kya	gya	sha	ja	cha	ja	nya	hya	bya	pya	mya	rya
yu	kyu	gyu	shu	ju	chu	jo	nyu	hyu	byu	pyu	myu	ryu
yo	kyo	gyo	sho	jo	cho	jo	nyo	hyo	byo	pyo	myo	ryo

Tabularised version of the Hepburn style of romanisation for yōon. Marked in red are changes compared to Kunrei style.

Contrary to Kunrei, Hepburn style uses macrons to represent long vowels, like in gojūon and yōon.

As you might have already realised, Hepburn style is not native. It was invented by US physician James C. Hepburn in 1867. Hepburn style does not follow a regular, systematic approach at romanising Japanese. Instead, it leverages English orthography and pronunciation to represent Japanese sounds. This means that shi_Hepburn represents the sound [ʃiː] instead of [ɕi]. Thus,

\[し \neq shi_{Hepburn} = shi_{English} \neq shi_{Spanish}.\] \[ほ \neq ho_{Hepburn} = ho_{English} \neq ho_{Spanish}.\]

The issue

A good romanisation system should be consistent, universally applicable, universally usable, and should faithfully and accurately reflect the source language. By using English as a crutch, Hepburn style not only does not provide an authentic insight into the Japanese language, it also essentially hinders non-English speakers the ability to precisely integrate this romanisation into their own language.

One might argue that English, being de facto lingua franca worldwide, is justified in being a support language when interfacing with other, non-Latin-based languages. For this particular case, however, the idea is quite daft and detrimental to the language being “helped”, as English is not even consistent with itself with regard to pronunciation, orthography, nor spelling.

Arabic has multiple romanisation methods, most of them not attached to any third language. In the cases where the romanisation style uses foreign orthography, like the case of Arabizi style, which is an informal romanisation method that leverages the orthography of either French or English, it uses numerals when the Latin alphabet is not enough to transcribe or fully convey parts of the language. Another example of a consistent, accurate, and universal romanisation system is pinyin, also recognised by ISO in ISO 7098 and also using its own notation set, independent of the pronunciation and orthography from a third language, to transcribe the inherent idiosyncrasies of Mandarin Chinese.

Leveraging a third language to romanise or transcribe any source language subordinates the latter to the former. The change in the official style of romanisation is, in a way, quite poetical, since it was General MacArthur who pushed for the use of Hepburn style after Japan became yet another colony of the US.

An example

The word for sheep is 羊 (ひつじ). In Kunrei style, it would be romanised as hituzi, while in Hepburn style, it would be romanised a hitsuji. It is entirely true that an anglophone would approximate the pronunciation of the original word better when reading hitsuji. Let us take now a middle-aged Spanish native person. This individual would natively pronounce the word as /it͡suxi/, which is very different from both the English pronunciation and the Japanese one.

“But wait”, one might say, “wouldn’t they get Kunrei style wrong as well?” The answer to this is “yes”. The main difference lies in the fact that, while both romanisation methods require further knowledge about the source language, Kunrei style does not require the user to rely on English as a middleware and can interact directly with the original language without the need of an intermediary. This is better in the long run for the source language, the user, and for learners of the source language, as the user is actively using and interacting with the original language.