Romaji Explained: Hepburn, Kunrei-shiki, and Nihon-shiki

Three Japanese romanization systems coexist because each was designed for a different reader. Hepburn helps English speakers approximate Japanese pronunciation. Kunrei-shiki gives the Japanese government phonemic regularity. Nihon-shiki helps linguists preserve historical kana distinctions.¹²³ On December 22, 2025, a Cabinet notification placed Modified Hepburn at the center of the official Japanese-government standard, ending the 71-year reign of the 1954 Kunrei-shiki framework.⁴⁵⁶

This guide is for two readers. If you are an absolute beginner who has seen the same word romanized as Shinjuku, Sinzyuku, and Sinjuku, you get the short answer immediately and the comparison table that follows it. If you are an intermediate learner researching the 2024 to 2025 policy revision, you get the full timeline, each system's design rationale, and a maintenance-flagged section on what the December 22, 2025 notification changed.⁴³⁶

Pick Modified Hepburn unless you have a specific reason not to

For anything a foreign reader will see, write Modified Hepburn with macrons (Tōkyō, sushi, Shinjuku). The exceptions are narrow: Japanese elementary-school worksheets used Kunrei-shiki through 2025; linguistic transcription uses Nihon-shiki; passport names follow a permissive Ministry of Foreign Affairs (MOFA) Hepburn variant of their own.⁷⁸³

The Short Answer

Use Modified Hepburn for anything a foreign reader will see

Modified Hepburn is the system the Japanese government uses on passports and in most international communications. It is also the system most widely used in English-language publications inside and outside Japan.³ The University of Tokyo Komaba style guide states the rule directly: "If you have no preference for a particular romanization system, we recommend that you use the Hepburn-based system."³

Three exceptions bend the rule. Japanese elementary schools taught Kunrei-shiki in the 3rd-grade rōmaji unit from 1954 until the December 22, 2025 Cabinet notification. Textbooks are scheduled to phase in Hepburn-based instruction starting in fiscal year 2026.⁵⁶ Linguistic transcription and historical-text editing use Nihon-shiki (ISO 3602 Strict) because it preserves kana distinctions that modern speakers no longer pronounce.⁹² The Ministry of Foreign Affairs passport romanization is its own Hepburn variant: macrons are not written, and applicants may use "OH" (since April 1, 2000) or "OO" / "OU" (since February 1, 2008) for long ō.⁷⁸

Why this question exists at all

The 1937 Cabinet Directive 「国語ノローマ字綴方統一ノ件」 was the first attempt to unify Japan's romanization. It settled a decades-long dispute between two camps: Hepburn-bloc supporters, who descended from James Curtis Hepburn's 1867 dictionary, and Japanese-nationalist phonemic-system supporters, who descended from Tanakadate Aikitsu's 1885 paper.^10111213

The 1937 directive was replaced by the 1954 Cabinet Notification No. 1 「ローマ字のつづり方」. That notification kept Kunrei-shiki on Table 1 (recommended) and listed Hepburn-style and Nihon-shiki-style spellings on Table 2 (permitted where established by practice).¹⁴⁵ The 1954 framework remained in force for 71 years. The Cabinet notification of December 22, 2025 abolished it and placed Modified Hepburn at the center of the new official standard.⁴⁶

The lineage of the three systems and the policy decisions that anointed each one as standard fits on a single timeline:

What Romaji Is (and Is Not)

Romaji is a transliteration tool, not a fourth Japanese script

Romaji (ローマ字, literally "Roman letters") is the practice of writing Japanese in the Latin alphabet. Fluent written Japanese mixes kanji, hiragana, and katakana. It uses romaji only for proper-noun signage, acronyms like NHK and JR, and foreign brand names.¹⁴³

The 1954 Cabinet Notification framed romaji as supplementary to the kana-and-kanji writing system. The 2024 ministerial consultation document restated this position. It noted that romaji's role has shifted from "writing Japanese in Latin letters domestically" to "conveying information to the international community."¹⁵¹⁴¹⁶

Where you actually see romaji in Japan

Romaji appears on station signs (JR and most private rail), road signs (governed by the Ministry of Land, Infrastructure, Transport and Tourism), passports (MOFA), business cards, branded product names, URLs, and personal names in international contexts.⁷¹⁷

Each surface follows its own house style

JR station signs follow a variant of Modified Hepburn; MOFA passports use a permissive Hepburn variant without macrons; academic publications generally follow ALA-LC Modified Hepburn or the Komaba / SWET style. As a result, three identical hiragana sequences can be romanized three different ways depending on where they appear.⁷³¹⁸¹⁹

The Three Systems

Hepburn (ヘボン式): the English-phonology system

Origin. James Curtis Hepburn's A Japanese and English Dictionary; with an English and Japanese Index first appeared in 1867. It was printed in Shanghai by the American Presbyterian Mission Press, with the imprint Yokohama / London, in 1,200 copies.¹¹ The system that English speakers now call "Traditional Hepburn" took its modern shape in the dictionary's third edition, published in 1886.²⁰¹

Modified Hepburn, also called Revised Hepburn or 修正ヘボン式, is a 1908 revision associated with educator Kanō Jigorō and the Romaji-Hirome-kai. It changed two Traditional Hepburn rules: syllabic ん is no longer rewritten as m before b/m/p (Shinbashi, not Shimbashi). An apostrophe disambiguates n followed by a vowel or y (hon'i, hon'yaku).¹

Long vowels. Modified Hepburn uses macrons (ā, ī, ū, ē, ō). The Komaba style guide accepts circumflex (â, î, û, ê, ô) as the typographic fallback when macrons cannot be set.³

Codification in foreign style guides. Modified Hepburn is the basis of the ALA-LC Japanese Romanization Table used by North American research libraries, adopted by the Library of Congress in 1989 and revised in 2012.¹¹⁹ The American National Standard "System for the Romanization of Japanese (Modified Hepburn)," ANSI Z39.11-1972, was withdrawn by ANSI in 1994; ANSI subsequently recommended Kunrei-shiki (ISO 3602) for catalog work, while ALA-LC retained Modified Hepburn for library cataloging.²¹¹⁹

The MOFA passport variant is not Modified Hepburn

The Ministry of Foreign Affairs uses its own Hepburn-based table. Macrons are not written, and long ō may be rendered "OH" (permitted from April 1, 2000) or "OO" / "OU" (permitted from February 1, 2008) for personal names. A bearer of the surname 大野 may register as Ono, Ohno, Oono, or Ouno. The choice is the bearer's and cannot be changed after registration.⁷⁸

Modified Hepburn is used today in international media, language textbooks aimed at non-Japanese audiences, JR station signs, road signs, passport name transcription, and the December 22, 2025 Cabinet notification standard.⁴⁷³¹⁷⁶

Kunrei-shiki (訓令式): the Japanese-government system

Origin. Kunrei-shiki begins with the 1937 Cabinet Directive No. 3 「国語ノローマ字綴方統一ノ件」. It was dated September 21, 1937 (昭和12年9月21日), and published in the 官報 (Official Gazette) the same day.¹⁰¹³

Re-codification. The 1954 Cabinet Notification No. 1 「ローマ字のつづり方」 was dated December 9, 1954 (昭和29年内閣告示第1号). It appeared alongside the parallel 内閣訓令第1号 「ローマ字のつづり方の実施について」 of the same date. Table 1 of the 1954 notification is the Kunrei-shiki set. Table 2 lists permitted alternative spellings drawn from Hepburn and Nihon-shiki for cases where customary use had diverged from Kunrei.¹⁴⁵

Design goal. Kunrei-shiki is a phonemic system that maps each kana column to a single Latin consonant: every t-row kana starts with t (ta, ti, tu, te, to), every s-row with s (sa, si, su, se, so), every h-row with h (ha, hi, hu, he, ho). This regularity is what the Komaba guide praises as "easier for native Japanese speakers to learn and more suitable for linguistic descriptions."³

International codification. ISO 3602:1989, Documentation: Romanization of Japanese (kana script), first edition published September 1, 1989, codifies Kunrei-shiki as the base ISO standard.⁹²¹

Kunrei was the school standard for 71 years and is now historical

Kunrei-shiki was in force from December 9, 1954 until its abolition by the Cabinet notification of December 22, 2025. Japanese elementary 3rd-grade rōmaji units used it for that entire span. Hepburn-based instruction phases in starting fiscal year 2026.⁴⁵⁶

Kunrei-shiki survives today in a residue of academic Japanese-linguistics publications and in libraries that follow ISO 3602.⁹³

Nihon-shiki (日本式): the linguistic-transcription system

Origin. Nihon-shiki traces to physicist Tanakadate Aikitsu (田中舘愛橘) and his 1885 paper 「理學協會雜誌ヲ羅馬字ニテ發兌スルノ發議及ヒ羅馬字用法意見」 in 理學協會雜誌 no. 16. The paper proposed a phonemic romanization built on the historical 50-on kana grid as an alternative to Hepburn's English-phonology system.¹²²¹³

Design goal. Nihon-shiki aims for a strict one-to-one mapping between the historical kana grid and Latin letters: every kana row gets one consonant, including the historical w-row (wa, wi, we, wo) and the k/g-rows with the now-extinct kwa / gwa contrast.²

Distinctions preserved that Hepburn and Kunrei merge. ぢ → di (vs Hepburn / Kunrei ji / zi), づ → du (vs zu), ゐ → wi (vs i), ゑ → we (vs e), くゎ → kwa (vs ka), ぐゎ → gwa (vs ga); the particle を → wo (vs Hepburn / Kunrei o).²

International codification. ISO 3602 Strict, defined as the strict subset of ISO 3602:1989 that preserves all historical kana distinctions.⁹²

Nihon-shiki is used today for linguistic transcription of historical Japanese, classical-text editing, reversible kana ↔ Latin conversion tools (corpus pipelines, OCR), and the rare publication that needs to preserve pre-modern orthography unambiguously.⁹²

How the Three Systems Diverge

The core divergence table

Each row below is a single fact about how one kana renders across the three systems.

Kana	Hepburn	Kunrei-shiki	Nihon-shiki	Notes
し	shi	si	si	Hepburn rewrites the kana because the actual sound /ɕi/ is closer to English "shi" than to "si."13
ち	chi	ti	ti	Hepburn writes "chi" to match the affricate /tɕi/; Kunrei keeps "t" for column regularity.13
つ	tsu	tu	tu	Hepburn writes "tsu" for the affricate /tsu/; Kunrei keeps "t."13
ふ	fu	hu	hu	Hepburn writes "fu" for the bilabial fricative /ɸu/; Kunrei keeps "h."13
じ	ji	zi	zi	All three systems agree that じ and ぢ are pronounced identically in modern Tokyo Japanese, but only Hepburn writes "ji."12
ぢ	ji	zi	di	Headline Nihon-shiki distinction: Nihon-shiki refuses the merger and keeps "di," reflecting historical /di/.2
づ	zu	zu	du	Parallel to ぢ: Nihon-shiki preserves "du" against Hepburn / Kunrei "zu."2
しゃ	sha	sya	sya	Hepburn writes the palatalized sound directly; Kunrei and Nihon-shiki build it from s + ya.123
ちゃ	cha	tya	tya	Same pattern: Hepburn writes the affricate, others build from t + ya.123
じゃ	ja	zya	zya	Hepburn writes the merged sound; Kunrei / Nihon-shiki build from z + ya.123
を (as particle)	o	o	wo	Hepburn and Kunrei follow the modern pronunciation /o/; Nihon-shiki preserves the historical "wo" form.1423
ん (before b/m/p)	n (Modified) / m (Traditional)	n	n	Modified Hepburn writes "n" in all positions (Shinbashi); Traditional Hepburn writes "m" before bilabials (Shimbashi); Kunrei and Nihon-shiki write "n" in all positions.13

The same sentence reads differently depending on which system the writer chose. Both romanized lines describe the same Japanese sentence:

寿司すしを食たべる。³
"I eat sushi." (Hepburn: Sushi o taberu. / Kunrei: Susi o taberu.)

Where Hepburn breaks from kana-column regularity

Hepburn rewrites a kana whenever its modern pronunciation diverges from what the column letter would predict: し is sh-, ち is ch-, つ is ts-, ふ is f-, じ is j-. The cost is that the s-row (sa, shi, su, se, so), the t-row (ta, chi, tsu, te, to), and the h-row (ha, hi, fu, he, ho) each contain a break in the column pattern. Kunrei keeps every column letter.³

The Komaba guide states the trade-off directly: Kunrei-shiki provides a closer one-to-one correspondence between Japanese kana and English letters. But English speakers who do not know Japanese can guess the actual sounds of these syllables more accurately from the Hepburn romanizations.³

H changes the consonant, K keeps the column

Hepburn rewrites whenever pronunciation has drifted (shi, chi, tsu, fu, ji); Kunrei keeps the column letter (si, ti, tu, hu, zi). The mnemonic captures the design philosophy and lets a learner predict the divergence without memorizing every row.³

Nihon-shiki goes one step further: it preserves the column letters even when modern pronunciation has merged distinctions away. That is why it keeps di for ぢ, du for づ, and wo for the particle を, regardless of how a contemporary Tokyo speaker actually pronounces them.²

Long vowels: the notation that fractures every system

Four conventions compete for the same long-vowel sound, and the choice depends entirely on the publication. The underlying kana patterns are described separately for hiragana and katakana. This section covers how each romanization system writes those patterns down.

Macron is the form preferred by Modified Hepburn, ALA-LC, and the December 22, 2025 Cabinet notification: long ō is written with a macron in Tōkyō, Kyōto, Ōsaka.⁴³¹⁹

Circumflex appears in Kunrei-shiki and Nihon-shiki publications and as a Hepburn typographic fallback: long ō appears as ô in Tôkyô. The Komaba guide accepts circumflex specifically as the Modified Hepburn fallback when macrons are not available.³

Doubled letter is the wāpuro / typing convention and the Cabinet notification's secondary permitted form: long ō appears as ou or oo in Toukyou or Tookyoo. The 1954 Cabinet Notification already accepted "oo" for native kanji-bound long vowels and "ou" as the alternative IME spelling. The 2025 notification continues to allow both macron and doubled-letter representations.⁴¹⁴⁵⁶

"OH" is the MOFA passport variant only. Long ō may be rendered "OH" on passports for personal names, effective April 1, 2000, restricted to the long-vowel reading of おお and おう. "OO" and "OU" became permitted alternatives from February 1, 2008.⁷⁸

東京とうきょうは大おおきい都市としです。³¹⁷
"Tokyo is a big city." (Hepburn: Tōkyō wa ōkii toshi desu. / Kunrei: Tôkyô wa ôkii tosi desu.)

A macron-less Tokyo is a different convention, not a typo

Reading material that prints Tokyo, Osaka, Kyoto without macrons is using a permissive Hepburn variant, most often MOFA passport style or an ASCII-fallback house style. It is not Modified Hepburn done wrong. A learner copying the spelling into academic writing should restore the macron: Tōkyō, Ōsaka, Kyōto.⁷⁸³

Two related rules round out long-vowel handling.

The Komaba "OO reset" rule: when the two vowels in a kanji-boundary combination would be represented by separate kanji, the macron is replaced by the doubled vowel. ばあい (場合) is romanized baai, not bāi. かつうら (勝浦) is Katsuura, not Katsūra. The rule keeps the kanji-morpheme boundary visible to the reader.³

The "ei" rule: all standard Hepburn variants spell けい / せい with "ei" (keizai, seifu, sensei), not "ē." The pronunciation is identical to a long ē, but Japanese phonology and orthography both treat the sequence as e + i rather than as a long vowel.³

Particle exceptions (は, へ, を)

All three systems agree on は as wa (topic marker) and へ as e (directional marker). These forms are written phonetically in romaji regardless of kana spelling.¹⁴³ The split is on を (object marker). Hepburn and Kunrei-shiki both write o, following the modern pronunciation /o/. Nihon-shiki preserves wo, following the historical kana.²³ The 1954 Cabinet Notification (Kunrei) and the 2025 revision (Hepburn) both codify the "o" form for the particle.⁴¹⁴⁵

本ほんを読よむ。¹⁴²
"I read a book." (Hepburn / Kunrei: Hon o yomu. / Nihon-shiki: Hon wo yomu.)

When Each System Is the Right Choice

A learner deciding which form to write rarely needs to memorize three full systems. In most cases, the audience and publication context settle the choice:

Default: Modified Hepburn for anything aimed at non-Japanese readers

The Komaba guide states the default position: Hepburn is the system most widely used in English-language publications in Japan and other countries, and it is the system recommended for academic writing.³ Long vowels take macrons (Tōkyō). Doubled vowels (Tookyoo, Toukyou) are acceptable when macrons cannot be rendered. "OH" is reserved for personal names whose bearer prefers it.⁷³ Syllabic ん takes "n" in all positions per Modified Hepburn (Shinbashi, not Shimbashi), with an apostrophe before vowels and y to disambiguate (hon'i, hon'yaku).¹³ For North American research libraries, ALA-LC Modified Hepburn is the cataloging-equivalent default.¹⁹

新橋しんばしで会あいましょう。¹
"Let's meet at Shinbashi." (Modified Hepburn: Shinbashi de aimashō. / Traditional Hepburn: Shimbashi de aimashō.)

Kunrei-shiki for Japanese-pedagogy contexts

Japanese elementary-school 3rd-grade rōmaji units used Kunrei-shiki from the 1954 notification until the December 22, 2025 Cabinet notification. Textbooks phase in Hepburn-based instruction starting in fiscal year 2026.¹⁴⁵⁶ Linguistic descriptions of Japanese that need to display column regularity (verb conjugation tables that align the t-row: tat-, tati, tatu, tate, tato) still benefit from Kunrei, because the morpheme boundary stays visible.³ The Komaba guide notes that this regularity makes Kunrei-shiki easier for native Japanese speakers to learn.³

Nihon-shiki for linguistic transcription

Nihon-shiki fits three use cases: editing classical or pre-modern Japanese where historical spelling matters; building reversible kana ↔ Latin pipelines (corpus tooling, OCR, lossless input methods); and academic phonological transcription that needs to preserve ぢ / じ and づ / ず.⁹² Nihon-shiki is the only one of the three systems that survives a round-trip transformation kana → Latin → kana without ambiguity for historical texts.²

鼻血はなぢが出でた。²
"I got a nosebleed." (Hepburn: Hanaji ga deta. / Kunrei: Hanazi ga deta. / Nihon-shiki: Hanadi ga deta.)

What about wāpuro (typing) romanization?

IME romaji input is a hybrid convention for typing Japanese. It accepts both Hepburn (shi, chi, tsu, fu) and Kunrei (si, ti, tu, hu) for kana entry, and uses doubled vowels to type long sounds (toukyou for とうきょう, tookyoo for native おお).³ The IME layer adds ASCII-only conventions: nn for syllabic ん (Keihinn), ltu or xtu for small っ standalone, and zk-prefixed key combinations for punctuation.³

Typing romaji is a writing-input convention, not a fourth system

The Komaba guide treats wāpuro as a writing-input convention rather than a publication-grade system. It has not been adopted officially by any major publications and is not suitable for academic writing. A learner should not romanize finished prose in wāpuro form.³

The December 2025 Cabinet Notification

What changed and what it means

The revision moved through five milestones, each tied to a specific date.

• January 2024. The Cultural Affairs Agency tabled a review proposal at the 文化審議会国語分科会 (Japanese Language Subcommittee of the Cultural Council). This began consideration of a revision to the 1954 notification.¹³
• May 14, 2024. Minister of Education Moriyama Masahito formally handed the consultation document (諮問) 「これからの時代におけるローマ字使用の在り方について」 to Shimaya Hiroyuki, Chair of the Cultural Council.¹⁵
• August 20, 2025 (令和7年8月20日). The Cultural Council delivered its formal recommendation 「改定ローマ字のつづり方（答申）」 to MEXT, with Chair Shimaya handing the document to Vice-Minister Nonaka at the second general assembly of the 25th Cultural Council.⁴²²
• December 16, 2025. The Cabinet decided to issue the new Cabinet notification, with publication scheduled for December 22, 2025.⁶
• December 22, 2025. The Cabinet notification abolished the 1954 Cabinet Notification No. 1 and Cabinet Directive No. 1. It established Modified Hepburn (with macron long-vowel notation) as the official Japanese-government romanization standard. Hepburn forms now official include shi, chi, tsu, fu, ji, sha, cha, ja.⁴¹⁴⁶

The same five milestones, on a single sequence diagram:

What the notification leaves to individual choice

Personal names and organization names: the 答申 (formal recommendation) explicitly preserves the bearer's existing romanization choice. A bearer of the surname 伊藤 may keep "Itoh," "Itou," or "Ito" rather than being forced into "Itō."⁴⁷⁸

Long-vowel notation: both macron (Tōkyō) and doubled letters (Tookyou, Tookyoo) remain permitted. The Cabinet notification accepts either.⁴⁶ Naturalized loanwords already established in English use (judo, matcha, sumo, Shimbashi as a station name) are not required to be rewritten.⁴¹⁵⁶ Scientific, artistic, and specialized field publications, and existing historical works, are explicitly exempted from forced retroactive changes.⁴⁶ Implementation in school textbooks is phased: the elementary 3rd-grade rōmaji unit teaches Hepburn-based forms starting in fiscal year 2026.⁶

The December 22, 2025 milestone is the maintenance trigger for this article

After December 22, 2025, "the Japanese government's romanization system" refers to Modified Hepburn (with macron long vowels) unless the text being cited is dated before that revision. Future maintenance should watch for later notifications, textbook adoptions, or MOFA passport-rule revisions.⁴⁶

What this article reflects going forward

The Kunrei-shiki section of this article is retained as historical context, not as a recommendation. The Nihon-shiki section is retained as a working description of ISO 3602 Strict for linguistic and archival users.⁹²

Good to know

Why your passport romaji is not the romaji you expected

The Ministry of Foreign Affairs passport convention is a Hepburn variant that omits macrons and permits "OH" for おお and "OO" / "OU" for おお or おう (Ohno, Itoh, Oono).⁷⁸ If you are choosing how to register your name on a passport, know that this variant is not Modified Hepburn and is not Kunrei. It is its own table, and the choice cannot be changed once registered.⁷⁸ A learner who writes Ōno in academic prose and Ohno on a passport is using both systems correctly, in their respective contexts.

Mixing Hepburn and Kunrei in the same document

Writing Tokyo in one paragraph and Tukiji in the next mixes Hepburn and Kunrei within one document. The Komaba style guide treats system consistency as a baseline requirement: pick one system per document and stay in it. The correct Hepburn pairing is Tokyo and Tsukiji. The correct Kunrei pairing is Tôkyô and Tukizi.³ An editor's copy-edit pass should flag mixed systems as an error, not as a stylistic choice.

Kunrei-shiki in foreign-audience contexts

Even before the 2025 revision, Hepburn was the de facto standard on passports, road signs, JR stations, and international communications.⁷³¹⁷¹⁶ A Kunrei spelling like Huzi-san for 富士山 will read as either pedantic or unfamiliar to a non-Japanese audience. Most foreign readers will not recognize it as the same mountain. Reserve Kunrei for materials whose audience is Japanese learners of romaji.

"OH"-style passport spellings outside their MOFA context

Spellings like Ohno and Itoh are permitted on Japanese passports for personal names only.⁷⁸ Using them for place names (Ohsaka instead of Ōsaka) is non-standard in every other Hepburn variant and reads as idiosyncratic.³ The MOFA "OH" rule does not generalize.

Hepburn was a missionary, not a linguist

James Curtis Hepburn (1815 to 1911) was an American Presbyterian medical missionary. His 1867 A Japanese and English Dictionary was printed in Shanghai by the American Presbyterian Mission Press and compiled over roughly eight years with the assistance of Japanese scholar Kishida Ginkō.¹¹ The romanization conventions used in the dictionary's headwords were a typographic device for English-speaking readers, not a designed standard. The system was formalized only in the third edition (1886). In Japanese, it was named ヘボン式 after the Japanese-era transcription of Hepburn's surname (ヘボン, Hebon).²⁰¹

Nihon-shiki was a nationalist phonology project, not a translation tool

Tanakadate Aikitsu (1856 to 1952), a physicist trained at the University of Tokyo and later in Glasgow, proposed Nihon-shiki in 1885 in the journal 理學協會雜誌. The founding aim was to eventually replace the kanji-kana system with romanized Japanese, so Japan could more easily compete with Western countries.¹²²¹³ This origin is why Nihon-shiki is the most regular of the three systems: it was designed for Japanese readers of romanized Japanese, not for English speakers approximating Japanese pronunciation.²

"Shinjuku, Sinzyuku, Sinzyuku" anchors the comparison table

The neighborhood 新宿 (しんじゅく) renders as Shinjuku in Hepburn and Sinzyuku in both Kunrei-shiki and Nihon-shiki. No ぢ / づ / ゐ / ゑ appears, so the divergence is only the s-row column break (し → shi vs si) and the s-row palatalization (じゅ → ju vs zyu).¹²³ Memorizing this one well-known place name anchors the entire divergence table without having to recall every row.

Why romaji has a sharp expiration date as a reading mode

Romaji is a transliteration tool for non-Japanese audiences and for input methods, not a fourth Japanese script.¹⁴³ The Komaba guide and the 2024 ministerial consultation both frame romaji as supplementary to kana and kanji. It is suited to a week-one pronunciation key, but it stops scaling once a learner can read hiragana.¹⁵¹⁴³ Lingering in romaji past the first few weeks delays the kana-to-kanji transition that fluent Japanese reading depends on.

See also

• Common Romaji Mistakes That Mislead Pronunciation
• The Japanese Consonant Inventory: Phonemes, Allophones, and the Kana Chart
• Long vs. Short Vowels in Japanese: The Distinction Beginners Miss
• Why "Tokyo" Is Two Syllables in English and Four Morae in Japanese: Loanwords as a Timing Drill
• Hiragana, Katakana, or Kanji First? A Beginner's Script Order

References

Footnotes

1. Wikipedia contributors. "Hepburn romanization." Wikipedia, The Free Encyclopedia. https://en.wikipedia.org/wiki/Hepburn_romanization (limitation: used only for cross-verified historical and variant summaries; primary sources cited where available). ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷

2. Wikipedia contributors. "Nihon-shiki romanization." Wikipedia, The Free Encyclopedia. https://en.wikipedia.org/wiki/Nihon-shiki_romanization (limitation: used only for system-comparison verification; primary sources cited where available). ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰ ↩²¹ ↩²² ↩²³ ↩²⁴

3. Department of English Language / Komaba Organization for Educational Development, The University of Tokyo, Komaba. Recommended System for Romanizing Japanese, v1 (2009-04). http://park.itc.u-tokyo.ac.jp/eigo/UT-Komaba-Romanization-of-Japanese-v1.pdf ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰ ↩²¹ ↩²² ↩²³ ↩²⁴ ↩²⁵ ↩²⁶ ↩²⁷ ↩²⁸ ↩²⁹ ↩³⁰ ↩³¹ ↩³² ↩³³ ↩³⁴ ↩³⁵ ↩³⁶ ↩³⁷ ↩³⁸ ↩³⁹ ↩⁴⁰ ↩⁴¹ ↩⁴² ↩⁴³ ↩⁴⁴ ↩⁴⁵ ↩⁴⁶

4. Agency for Cultural Affairs (文化庁). 「改定ローマ字のつづり方（答申）」, presented to the Minister of Education on August 20, 2025 (令和7年8月20日). https://www.bunka.go.jp/seisaku/bunkashingikai/sokai/pdf/94261201_01.pdf ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵

5. Cabinet of Japan (内閣). 「ローマ字のつづり方」, Cabinet Notification No. 1 of 1954 (内閣告示第1号, 昭和29年12月9日), with Cabinet Directive No. 1 of the same date (内閣訓令第1号「ローマ字のつづり方の実施について」). Wikisource reproduction: https://ja.wikisource.org/wiki/ローマ字のつづり方_(昭和29年内閣告示第1号) ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸

6. 時事通信 (Jiji Press). 「ローマ字新表記、22日告示 約70年ぶり改定、ヘボン式基本に」, December 16, 2025. https://www.jiji.com/jc/article?k=2025121600387&g=soc ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵

7. Ministry of Foreign Affairs of Japan (外務省). 「ヘボン式ローマ字綴方表」, official passport romanization table. https://www.ezairyu.mofa.go.jp/passport/hebon.html ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴

8. Tokyo Metropolitan Government, Citizens, Culture and Sports Bureau (東京都生活文化スポーツ局). 「氏名の表記（ヘボン式ローマ字等）」, passport-application guidance documenting MOFA non-Hepburn allowances including "OH" for long ō and "OO/OU" for おお/おう (effective 2000-04-01 and 2008-02-01 respectively). https://www.seikatubunka.metro.tokyo.lg.jp/passport/documents/0000000485 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹

9. International Organization for Standardization. ISO 3602:1989, Documentation: Romanization of Japanese (kana script). First edition 1989-09-01. https://www.iso.org/standard/9029.html ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

10. Cabinet of Japan (内閣). 「国語ノローマ字綴方統一ノ件」, Cabinet Directive No. 3 of 1937 (内閣訓令第3号, 昭和12年9月21日), published in 官報 (Official Gazette), September 21, 1937. Archival reproduction: https://green.adam.ne.jp/roomazi/tuzurikata.html ↩ ↩²

11. Hepburn, James Curtis. A Japanese and English Dictionary; with an English and Japanese Index. First edition, Shanghai: American Presbyterian Mission Press, 1867 (imprint: Yokohama / London). Internet Archive: https://archive.org/details/japaneseenglishd00hepb ↩ ↩² ↩³

12. 田中舘愛橘 (Tanakadate Aikitsu). 「理學協會雜誌ヲ羅馬字ニテ發兌スルノ發議及ヒ羅馬字用法意見」 (Proposal to publish the Journal of the Science Association in romaji, and an opinion on the use of romaji). 理學協會雜誌 (Rigaku Kyōkai Zasshi), no. 16, 1885. Bibliographic reference catalogued at 田中舘愛橘記念科学館: https://www.nbsk.or.jp/tanakadate/業績コーナー/ ↩ ↩² ↩³

13. 文化審議会国語分科会 (Cultural Council, Japanese Language Subcommittee). 「ローマ字に関する国語施策関係年表」, 国語課題小委員会参考資料, 令和4年7月19日. https://www.bunka.go.jp/seisaku/bunkashingikai/kokugo/kokugo_kadai/iinkai_52/pdf/93733501_10.pdf ↩ ↩² ↩³ ↩⁴ ↩⁵

14. Agency for Cultural Affairs (文化庁). 国語施策情報・内閣告示・内閣訓令・「ローマ字のつづり方」(commentary and full text of the 1954 Cabinet Notification No. 1 / 昭和29年内閣告示第1号 and the Cabinet Directive of the same date). https://www.bunka.go.jp/kokugo_nihongo/sisaku/joho/joho/kijun/naikaku/roma/index.html ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³

15. Ministry of Education, Culture, Sports, Science and Technology (MEXT / 文部科学省). 「ローマ字の使用の在り方について盛山大臣から島谷文化審議会会長へ諮問文を手交」 (May 14, 2024 consultation hand-off, titled 「これからの時代におけるローマ字使用の在り方について」). https://www.mext.go.jp/b_menu/activity/detail/2024/20240514.html ↩ ↩² ↩³ ↩⁴

16. Nippon.com editorial staff. "'Sushi' Or 'Susi'? Japan to Rethink Longstanding Romanization Rules." Nippon.com, June 13, 2024. https://www.nippon.com/en/japan-data/h02005/ ↩ ↩²

17. Nippon.com editorial staff. "Akashi or Akasi? Hepburn Most Established of Japan's Different 'Rōmaji' Systems." Nippon.com, November 2, 2022. https://www.nippon.com/en/japan-data/h01483/ ↩ ↩² ↩³ ↩⁴

18. Society of Writers, Editors and Translators (SWET). "Hepburn System Romanization, Macrons, and Flexibility." https://swet.jp/news/article/hepburn_system_romanization_macrons_and_flexibility ↩

19. Library of Congress and American Library Association. ALA-LC Romanization Table: Japanese (2022 version). https://www.loc.gov/catdir/cpso/romanization/japanese.pdf ↩ ↩² ↩³ ↩⁴ ↩⁵

20. Hepburn, James Curtis. A Japanese-English and English-Japanese Dictionary. Third edition, Tokyo: Z. P. Maruya & Co., 1886. Internet Archive: https://archive.org/details/japaneseenglishe00hepb_0 ↩ ↩²

21. Wikipedia contributors. "Kunrei-shiki romanization." Wikipedia, The Free Encyclopedia. https://en.wikipedia.org/wiki/Kunrei-shiki_romanization (limitation: used only for cross-verification of ordinance numbers already pinned to government primary sources). ↩ ↩²

22. Ministry of Education, Culture, Sports, Science and Technology (MEXT / 文部科学省). 「文化審議会総会において、島谷文化審議会会長から野中副大臣へ「改定ローマ字のつづり方（答申）」を手交」 (August 20, 2025 handover of the 答申). https://www.mext.go.jp/b_menu/activity/detail/2025/20250820_2.html ↩