Japanese appears in Roman letters everywhere: passports, maps, railway signs, academic catalogs, brand names, language textbooks, websites, and keyboards.

The Japanese capital is commonly written Tokyo, but a fully marked form may appear as Tōkyō. The kana し may be represented as shi or si. The word ふじ may appear as Fuji or Huzi, while つづく can become tsuzuku, tuzuku, or tuduku.

These differences come from competing Japanese romanization systems.

The three most important are:

• Hepburn romanization
• Kunrei-shiki romanization
• Nihon-shiki romanization

Hepburn generally favors spellings that help international readers approximate modern Japanese pronunciation. Kunrei-shiki emphasizes the regular structure of the Japanese syllabary while reflecting modern standard pronunciation. Nihon-shiki follows kana distinctions even more strictly, preserving contrasts that are no longer maintained in ordinary modern speech.

Japan’s official policy changed significantly on December 22, 2025. The government replaced its 1954 guidance with a new principal table using familiar Hepburn-like forms such as shi, chi, tsu, fu, and ji. However, the government explicitly states that the new rules are not simply an adoption of every convention traditionally called Hepburn. (Bunka)

Quick Comparison

The most visible difference is that Hepburn uses pronunciation-oriented spellings such as shi, chi, and tsu, while Kunrei-shiki and Nihon-shiki use the more regular forms si, ti, and tu.

The deeper difference appears with kana such as じ and ぢ or ず and づ. Modern standard Japanese normally pronounces each pair alike. Hepburn merges them as ji and zu. Kunrei-shiki also merges them, but as zi and zu. Nihon-shiki preserves the written distinctions with zi/di and zu/du. Japan’s 2025 official comparison table displays these relationships directly.

What Is Japanese Romanization?

Japanese romanization is the representation of Japanese words with the Latin alphabet. In Japanese, Roman letters are called rōmaji.

Romanization can represent text originally written with:

• Hiragana
• Katakana
• Kanji
• A mixture of Japanese scripts

Kana can often be converted through a table of correspondences, but kanji must first be assigned the correct Japanese reading.

For example, 日本 can be read as:

• Nihon
• Nippon

The correct form depends on context, established usage, and sometimes the preference of an institution or speaker. The ALA-LC Japanese Romanization Table advises catalogers to use contextual evidence and recognized dictionaries because Japanese script does not always have a simple one-to-one relationship with one romanized reading. (The Library of Congress)

Romanization is not translation. Writing 東京 as Tōkyō represents the Japanese name in Latin letters. Translating 東京 literally as eastern capital is a different process.

The Current Official Position in Japan

For decades, Japan’s official guidance was associated primarily with Kunrei-shiki. The 1954 Cabinet Notification placed Kunrei-style spellings in its principal table, while allowing Hepburn and certain Nihon-shiki forms in a secondary table for international relations and established conventions.

That arrangement ended on December 22, 2025.

The new Cabinet Notification abolished the 1954 notice and introduced one principal table for representing modern Japanese in ordinary public life. That table uses:

• shi, not si
• chi, not ti
• tsu, not tu
• fu, not hu
• ji, not zi
• sha, not sya
• cha, not tya
• ja, not zya

The government adopted these forms because Hepburn-like spellings had become much more widely used in passports, road signs, guidance displays, and ordinary social life than the old principal Kunrei-style system.

However, the official Q&A cautions against describing the change as complete “unification under Hepburn.” The new table is Hepburn-based, but it differs from some traditional Hepburn conventions in the treatment of the moraic nasal, doubled consonants, and long vowels. Kunrei-shiki and Nihon-shiki may also continue to be used in specialist fields or individual contexts.

The new guidance applies broadly to modern Japanese in areas such as laws, government documents, newspapers, magazines, and broadcasting. It does not attempt to override all specialist conventions, personal spellings, historical documents, or established institutional names.

What Is Hepburn Romanization?

Hepburn romanization is a family of systems derived from the work of the American physician and missionary James Curtis Hepburn. The method was developed during the late Edo and early Meiji periods and became highly influential in dictionaries and communication with Western-language readers. Japan’s Cultural Affairs Agency describes Hepburn as a broad family rather than one completely uniform set of rules.

Hepburn is generally pronunciation-oriented. It uses Latin-letter combinations that readers familiar with English-influenced spelling may find more intuitive.

Examples include:

Words commonly written in Hepburn-style romanization include:

These spellings are widely recognizable internationally, but they are not perfectly reversible.

Why Hepburn Is Easier for Many International Readers

Hepburn’s shi, chi, and tsu provide stronger pronunciation cues for many readers than the Kunrei forms si, ti, and tu.

For example:

• shi suggests that し is not an ordinary si syllable.
• chi signals that ち differs from an ordinary ti.
• tsu shows the consonant sequence associated with つ.
• fu provides a closer approximation than hu for many readers.

These are still conventions rather than exact phonetic transcriptions. Japanese r, f, sh, vowels, and other sounds do not necessarily match their closest English equivalents exactly.

Why Hepburn Is Not Fully Reversible

Hepburn merges several kana distinctions:

• じ and ぢ both become ji
• ず and づ both become zu
• じゃ and ぢゃ both become ja

Once the original kana are removed, the Latin form alone may not reveal whether ji came from じ or ぢ.

Unicode’s transliteration guidance identifies this ambiguity and uses modified forms such as dji and dzu in one transform when greater reversibility is needed. (Unicode CLDR)

There Is More Than One Hepburn Variant

“Hepburn” can refer to several related conventions.

They may differ in their treatment of:

• ん before b, m, and p
• Small っ before ch
• Long vowels
• Apostrophes
• Word division
• Established names
• Foreign sounds written in katakana

The most relevant variants include traditional Hepburn, modified Hepburn, passport Hepburn, library Hepburn, and Japan’s new 2025 Hepburn-based public guidance.

A tool should therefore label its result more precisely than simply “Hepburn.”

What Is Kunrei-shiki?

Kunrei-shiki, or the Cabinet-ordinance system, is a systematic romanization method largely based on Nihon-shiki but adjusted to reflect modern standard Japanese pronunciation.

Its name comes from the 1937 Cabinet Instruction that established a unified government romanization scheme. The old 1954 Cabinet Notification later placed this approach in its primary table.

Kunrei-shiki represents the Japanese syllabary through highly regular consonant patterns.

Consider the s, t, and h rows:

The pattern is visually regular:

• Every syllable in the s row begins with s
• Every syllable in the t row begins with t
• Every syllable in the h row begins with h

This makes the structure of the kana table more apparent than Hepburn’s shi, chi, tsu, and fu.

Examples in Kunrei-shiki

Kunrei-shiki is more regular, but readers unfamiliar with the system may pronounce si, ti, and tu according to the spelling habits of their own languages.

Kunrei-shiki and Modern Pronunciation

Kunrei-shiki preserves the regular kana-row structure but does not preserve every kana distinction.

In contemporary standard Japanese, じ and ぢ are generally treated as having the same pronunciation. Kunrei-shiki therefore writes both as zi. Likewise, ず and づ both become zu.

For example:

This is one of the main differences between Kunrei-shiki and Nihon-shiki.

Kunrei-shiki and ISO 3602

ISO 3602:1989 establishes an international standard for the romanization of present-day Japanese written in kana. It reflects the earlier systematic approach associated with Japan’s former official principal table. ISO continues to list ISO 3602:1989, while the Japanese government’s 2026 Q&A states that reconsideration of the international standard is expected following the 2025 national revision. (ISO)

The national policy and the currently listed international standard are therefore not fully aligned.

What Is Nihon-shiki?

Nihon-shiki, or Nippon-shiki, is the most structurally conservative of the three major systems.

It was developed in the late nineteenth century by Japanese scholars including Tanakadate Aikitsu. Its purpose was to provide a regular representation of the Japanese kana system rather than make Japanese immediately pronounceable to readers of English.

Nihon-shiki forms the historical foundation from which Kunrei-shiki developed.

Structural Regularity

Like Kunrei-shiki, Nihon-shiki uses:

• si for し
• ti for ち
• tu for つ
• hu for ふ
• sya for しゃ
• tya for ちゃ
• zya for じゃ

Its distinctive feature is that it preserves separate kana spellings even where modern pronunciation no longer distinguishes them.

This makes Nihon-shiki more useful for identifying the original kana.

Example: かなづかい

The word かなづかい contains づ.

It may be represented as:

• Hepburn: kanazukai
• Kunrei-shiki: kanazukai
• Nihon-shiki: kanadukai

Nihon-shiki’s du preserves the fact that the word is written with づ rather than ず. Kunrei-shiki writes both kana as zu because they are normally pronounced alike in modern standard Japanese. Japan’s official historical explanation identifies this treatment of the “four kana” distinctions as a defining characteristic of Nihon-shiki.

Nihon-shiki and Historical Kana

Nihon-shiki can also preserve older distinctions such as:

• を as wo
• Historical くわ as kwa
• Historical ぐわ as gwa

The 2025 official comparison table retains these forms only as historical points of comparison. The new principal system uses o, ka, and ga for modern Japanese.

Hepburn, Kunrei-shiki and Nihon-shiki Compared

Basic Kana Differences

The new national table resembles Hepburn most visibly, but its supplementary rules create a distinct modern government standard.

Word Examples

The Shimbashi/Shinbashi variation demonstrates why “Hepburn” needs a variant label. Traditional Hepburn commonly changes ん to m before b, m, or p, while Japan’s 2025 principal rules consistently use n, producing Shinbashi. (Unicode CLDR)

The Moraic Nasal ん

The kana ん represents a moraic nasal whose exact pronunciation changes according to the surrounding sounds.

Romanization systems differ in whether the Latin spelling should reflect that phonetic variation.

Traditional Hepburn

Traditional Hepburn commonly changes n to m before:

• b
• m
• p

Examples include:

Japan’s passport guidance continues to prescribe m before b, m, and p for ordinary Hepburn passport forms. (Ezairyu)

New 2025 Japanese Main Table

The 2025 principal table writes ん consistently as n:

• あんまん → anman
• 乾杯 → kanpai
• 銀座 → Ginza
• 新聞 → shinbun

This produces a more stable relationship between kana and Roman letters than the traditional m substitution.

Consequently, passport Hepburn and the new general government guidance do not always produce identical results.

Small っ and Doubled Consonants

Small っ, called sokuon, represents a following doubled or geminated consonant.

Most systems repeat the initial consonant:

The treatment before ch differs among conventions.

Traditional or Library Hepburn

Some modified Hepburn systems place t before ch:

• みっちゃく → mitchaku
• ねっちゅう → netchū
• しゅっちょう → shutchō

The ALA-LC Japanese table uses this convention. (The Library of Congress)

New 2025 Main Table

The new Japanese guidance doubles the first letter of a two-letter consonant group:

• 日直 → nicchoku
• 雑誌 → zasshi

Thus, small っ before ch becomes cch, not tch, under the 2025 principal rules.

Passport Hepburn

Passport rules use tch before a chi syllable:

• ハッチョウ → HATCHO
• ホッチ → HOTCHI

This is another case in which “Hepburn” varies by institution. (Ezairyu)

Long Vowels

Long vowels are among the least consistently represented features of Japanese romanization.

Common methods include:

• Macrons: Tōkyō
• Circumflexes: Tôkyô
• Doubled vowels: Toukyou
• Kana-based vowel sequences: Tokyo, Toukyou, or other forms depending on policy
• Complete omission of the length distinction: Tokyo

Macrons

A macron is a horizontal mark above a vowel:

• ā
• ī
• ū
• ē
• ō

Examples include:

• 東京 → Tōkyō
• 柔道 → jūdō
• 大阪 → Ōsaka
• 京都 → Kyōto

Macrons preserve vowel length compactly.

The ALA-LC Japanese table uses macrons in forms such as Nyū Yōku, rōdō, and jōhō, subject to detailed morphological and spelling rules. (The Library of Congress)

New 2025 Options

Japan’s new official guidance allows two major approaches.

A long vowel may be represented with a macron:

• 東北 → Tōhoku
• 房総 → Bōsō
• 風流 → fūryū

It may also be represented by writing the vowels according to modern kana spelling:

• 東北 → Touhoku
• 房総 → Bousou
• 風流 → fuuryuu

A circumflex may substitute for a macron when necessary.

Passport Names

Japanese passport guidance generally omits o and u used to indicate long vowels in standard Hepburn forms:

• イトウ → ITO
• ユウキ → YUKI
• ヨウコ → YOKO
• オオノ → ONO

Certain approved alternatives may be available, but official identity spelling should always be taken directly from the document rather than regenerated automatically. (Ezairyu)

Established Unmarked Forms

Many internationally established names omit macrons:

• Tokyo
• Osaka
• Kyoto
• judo
• sumo
• Shinto

The 2025 guidance does not demand immediate replacement of all established spellings. Its examples explicitly contrast established forms such as Tokyo and judo with the fully marked forms Tōkyō and jūdō.

Apostrophes and Syllable Boundaries

An apostrophe helps distinguish the moraic ん from a following vowel or y syllable.

Examples include:

• 単位 → tan’i
• 船員 → sen’in
• 翻訳 → hon’yaku
• 信越 → Shin’etsu

Without the apostrophe, readers or software may divide the Roman letters incorrectly.

Both Japan’s 2025 guidance and ALA-LC library rules use apostrophes for this purpose.

Grammatical Particles

Some Japanese grammatical particles are spelled differently from their ordinary kana pronunciation or kana-row romanization.

For example:

• 私は学生です。
  Watashi wa gakusei desu.
• 東京へ行く。
  Tōkyō e iku.

The 2025 government rules specify wa, e, and o for these particles. The ALA-LC chart likewise accounts for the changed readings of は and へ when they function as grammatical particles.

Kanji Makes Romanization More Complicated

Kana tables do not solve every Japanese romanization problem because Japanese commonly uses kanji.

A kanji may have:

• Several Sino-Japanese readings
• One or more native Japanese readings
• A specialized personal-name reading
• An irregular compound reading
• A historical reading
• A reading selected by an author or institution

Examples include:

Romanization must therefore follow the word’s actual reading, not merely replace each kanji with one fixed Latin equivalent.

The ALA-LC system describes Japanese catalog romanization as a contextual process rather than a pronunciation guide. When readings are uncertain, catalogers may use dictionary evidence and add alternative readings as variant access points. (The Library of Congress)

Japanese Romanization in Library Catalogs

The ALA-LC Japanese Romanization Table serves a specialized bibliographic purpose.

Its goals include:

• Consistent catalog records
• Search and access points
• Standardized word division
• Treatment of titles and proper names
• Representation of historical spelling
• Preservation of useful distinctions

The current table is the 2022 version. It states that its romanized text aims to mirror Japanese script and readings as closely as practical, but it is not intended as a general pronunciation guide. (The Library of Congress)

ALA-LC forms can differ from public or learner-oriented romanization because the table applies detailed rules for:

• Long vowels
• Morpheme boundaries
• Historical kana
• Word division
• Capitalization
• Loanwords
• Variant readings

For example, a katakana word may be romanized as it actually appears in the item, even when its spelling differs from the modern usual form. (The Library of Congress)

Romanization and Japanese Keyboard Input Are Not the Same

Typing Japanese with Latin letters does not necessarily mean following a formal romanization standard.

A Japanese input method may accept several keystroke sequences:

• shi or si for し
• chi or ti for ち
• tsu or tu for つ
• ji, zi, or sometimes di for related kana
• wo for を

These are input commands used to produce kana and kanji. They are not automatically the form that should be displayed as romanized Japanese.

Japan’s Cultural Affairs Agency explicitly states that the 2025 spelling revision does not change computer input methods. Roman-letter input is a technical mechanism for entering Japanese script and is separate from writing Japanese itself in Roman letters.

A romanization converter should therefore distinguish:

• Formal romanized output
• Kana-to-Roman conversion
• Roman-keyboard input sequences
• Search normalization
• ASCII aliases

Which System Is Most Reversible?

Of the three principal systems, Nihon-shiki generally preserves the strongest correspondence with kana.

For example:

Hepburn loses the distinction first. Kunrei retains the regular row-based consonants but still merges kana that are pronounced alike. Nihon-shiki preserves the original spelling distinction.

However, even Nihon-shiki does not make all Japanese text automatically reversible because:

• Kanji readings must be determined before romanization.
• Word boundaries are not explicit in ordinary Japanese writing.
• Historical and contextual readings can vary.
• Proper names may have unpredictable readings.
• Romanization may omit punctuation or script information.

The original Japanese text should always be preserved.

Which System Is Easiest to Pronounce?

For many international users, Hepburn-like spelling is the most accessible because of forms such as:

• shi
• chi
• tsu
• fu
• sha
• cha
• ja

This accessibility helped Hepburn-based spellings become common in signs, passports, names, maps, and international communication.

However, Hepburn is not a universal phonetic guide. Its apparent readability is partly dependent on the reader’s language. A Latin spelling that seems intuitive to an English speaker may not be interpreted identically by a Spanish, Indonesian, German, or French speaker.

Audio or IPA should be provided when precise pronunciation matters.

Which System Is Most Systematic?

Kunrei-shiki and Nihon-shiki show the structure of the Japanese kana table more clearly.

Compare:

For linguistic analysis, kana organization, or certain forms of language instruction, this regularity can be valuable.

The Japanese government’s 2025 review acknowledged that Kunrei-shiki and Nihon-shiki have important systematic and structural qualities even though they are now rarely used in general public life.

Choosing the Right Japanese Romanization System

For General Websites

A practical public website should normally use familiar Hepburn-based forms:

• Tokyo
• Osaka
• sushi
• tsunami
• Fujisan

Macrons can be retained in reference content:

• Tōkyō
• Ōsaka
• jūdō
• shōgun

Search aliases should include unmarked forms.

For Academic Work

State the convention explicitly.

For example:

Japanese terms are romanized using modified Hepburn with macrons, except for established names and quoted historical spellings.

Do not switch silently between shi/si, chi/ti, or tsu/tu.

For Personal Names

Use the person’s official or preferred form.

A mechanical converter may produce a form that differs from:

• Passport spelling
• Professional identity
• Family convention
• Institutional record
• Established international usage

The 2025 government guidance specifically says that the wishes of individuals and organizations should be respected when writing personal names and organization names.

Best Practices for a Japanese Romanization Tool

A dependable tool should show more than one unlabeled result.

Preserve the Original Japanese

Display the original kanji and kana.

Identify the Reading

For kanji, show the kana reading used to generate the result.

Example:

Name the Exact Variant

Use labels such as:

• Japan 2025 principal table
• Traditional Hepburn
• Modified Hepburn
• Passport Hepburn
• ALA-LC Japanese 2022
• Kunrei-shiki
• Nihon-shiki

Separate Long-Vowel Options

Show:

• Marked form: Tōkyō
• Kana-based form: Toukyou
• Established form: Tokyo

Warn About Lost Distinctions

A Hepburn result such as ji may correspond to じ or ぢ.

Keep Keyboard Input Separate

Do not present IME keystrokes as though they were necessarily standard publication spellings.

Support Search Variants

A search index may connect:

• Tōkyō
• Tokyo
• Toukyou
• とうきょう
• トウキョウ
• 東京

Record the Standard Version

Japan’s 2025 policy change demonstrates why romanization metadata should include the authority, date, and variant.

Common Misconceptions

Hepburn Is One Fixed System

Hepburn is a family of related conventions. Passport, library, traditional, modified, and 2025 government forms can differ.

Kunrei-shiki Was Always the Most Common System in Japan

It was the principal system in the former official guidance, but the Japanese government concluded that it had not become established in ordinary social use. Hepburn-like spellings were much more common in practice.

The 2025 Reform Banned Kunrei-shiki

It did not. The new principal table applies to general modern Japanese, but the official Q&A says Kunrei-shiki and Nihon-shiki may continue where their specific functions are useful.

Nihon-shiki and Kunrei-shiki Are Identical

They share many spellings, but Nihon-shiki preserves kana distinctions such as di/du that Kunrei-shiki merges as zi/zu.

Tokyo Is the Fully Precise Romanization

Tokyo is an established international spelling. A vowel-length-marked form is Tōkyō, while Japan’s new guidance also permits Toukyou.

Passport Hepburn and General Hepburn Are Identical

Passport rules have their own treatment of long vowels, ん, and doubled consonants. The exact passport spelling should not be inferred from a generic Hepburn converter. (Ezairyu)

Romanization Can Be Generated Directly from Every Kanji

Kanji can have multiple readings. The intended word, context, or personal-name reading must be known.

Romaji Keyboard Input Defines Correct Romanization

Keyboard sequences are input commands. Formal romanization is a separate publication and language-policy issue.

Frequently Asked Questions

What is the official Japanese romanization system now?

Since December 22, 2025, Japan’s principal public guidance uses one Hepburn-based table with spellings such as shi, chi, tsu, fu, and ji. The government does not describe it as adopting every traditional Hepburn rule. (Bunka)

What is the difference between Hepburn and Kunrei-shiki?

Hepburn uses pronunciation-oriented forms such as shi, chi, and tsu. Kunrei-shiki uses structurally regular forms such as si, ti, and tu.

What is the difference between Kunrei-shiki and Nihon-shiki?

Kunrei-shiki follows modern pronunciation by merging じ with ぢ and ず with づ. Nihon-shiki preserves the kana distinctions as zi/di and zu/du.

Is Nihon-shiki more accurate?

It is more accurate for identifying certain original kana spellings. It is not necessarily the most helpful system for modern pronunciation or international communication.

Why is Fuji written Huzi in Kunrei-shiki?

The kana ふ belongs to the h row. Kunrei-shiki represents the row regularly as ha, hi, hu, he, ho, while Hepburn uses fu to provide a closer pronunciation cue.

Why is tsunami not written tunami?

Tsunami follows Hepburn. Kunrei-shiki and Nihon-shiki would use tunami.

Is Tokyo or Tōkyō correct?

Both can be appropriate. Tōkyō explicitly represents the long vowels. Tokyo is the established international form. The 2025 rules also allow Toukyou as a vowel-doubling alternative.

Why is Shimbashi sometimes written Shinbashi?

Traditional Hepburn commonly changes ん to m before b, producing Shimbashi. Japan’s 2025 table keeps n, producing Shinbashi.

Should Japanese learners use Hepburn?

Hepburn is useful for beginning pronunciation guidance, but learners should move to hiragana and katakana early because romanization cannot represent every aspect of Japanese writing or pronunciation independently.

Does ISO still use Kunrei-shiki?

ISO 3602:1989 remains the listed international standard for Japanese kana romanization. Japan’s Cultural Affairs Agency says consideration of how the international standard should respond to the 2025 national revision is expected. (ISO)

Which system does the Library of Congress use?

The ALA-LC Japanese Romanization Table uses a detailed bibliographic scheme based on modified Hepburn principles, together with its own rules for long vowels, word division, capitalization, historical spellings, and variant readings. Its current published table is the 2022 version. (The Library of Congress)

Conclusion

Hepburn, Kunrei-shiki, and Nihon-shiki represent three different answers to the same question: what should Japanese romanization preserve?

Hepburn prioritizes recognizable pronunciation for international readers. Its familiar forms include shi, chi, tsu, fu, and ji.

Kunrei-shiki prioritizes the regular organization of modern Japanese kana. It uses forms such as si, ti, tu, hu, and zi while merging kana that are pronounced alike.

Nihon-shiki preserves the written kana system more strictly. It distinguishes forms such as zi/di, zu/du, and o/wo, making it the most structurally reversible of the three.

Japan’s 2025 reform shifted general public guidance toward Hepburn-like spellings that had already become dominant in actual use. Yet it did not eliminate Kunrei-shiki, Nihon-shiki, passport conventions, library standards, or established personal and institutional spellings.

A reliable Japanese romanization reference should therefore show:

• The original Japanese script
• The kana reading
• The exact romanization system
• The treatment of long vowels
• The handling of ん and small っ
• Whether distinctions have been merged
• Alternative historical or official forms
• The authority and version behind the result

The purpose of romanization is not to identify one spelling as correct in every context. It is to explain what each spelling represents—and why shi, si, and other forms can all be systematic answers under different rules.