Skip to main content

Hiragana, Katakana, or Kanji First? A Beginner's Script Order

For absolute beginners, the hiragana, katakana, or kanji first question has one durable answer: hiragana, then katakana, then kanji learned alongside vocabulary. Use rōmaji only as a brief pronunciation key in the opening week or two.1 This article helps day-one learners decide where to put the first hour of study.

Overview

Japanese is written in three native scripts running together in the same sentence: hiragana, katakana, and kanji.2 Rōmaji is a transliteration system used for input and signage, not a fourth script you read full books in.3

The order in which beginners pick these up is not just a matter of taste. It follows from what each script does in a sentence, what mainstream textbooks expect you to read by each lesson, and what the language's phonology demands.12

The Short Answer

The four-step order at a glance

The recommended sequence, in order:

  1. Hiragana first. Particles and verb endings are written in hiragana, so a learner cannot parse a sentence without it.2
  2. Katakana second. Same sound system as hiragana, different shapes, used for loanwords, onomatopoeia, scientific terms, and emphasis.4
  3. Kanji third, and always with vocabulary. A kanji has multiple readings selected by lexical context; the readings only stabilize when you learn the words that use them.42
  4. Rōmaji as a pronunciation aid only. A bridge for the first week or two while you map sounds to hiragana shapes, then dropped.1

Genki I, the most widely adopted university-level beginner course, drills hiragana in Lesson 1, katakana in Lesson 2, and only begins introducing kanji from Lesson 3, after which rōmaji glosses are removed from the main text.1 Japanese elementary education mirrors the same order. The 学年別漢字配当表 (kyōiku list, 1,026 characters) is taught across six grades in vocabulary context, not as isolated character drill.5

Rōmaji is one tool with two jobs

Rōmaji as a reading aid for absolute beginners has a brief expiry date. Rōmaji as an input method (typing haru on a QWERTY keyboard and watching the IME convert it to はる) is permanent and standard.3 The two uses are easy to confuse, so separate them from day one.

Time budget per step

The U.S. Foreign Service Institute classifies Japanese in its highest difficulty tier, the Category V "super-hard" group. It estimates roughly 2,200 class hours (about 88 weeks) to general professional proficiency for adult English speakers.6 The first-month script-order question is only the very front of that curve. Getting the order right matters more than the exact hour count.

The Genki self-study material assumes a daily cadence of roughly fifteen to thirty minutes for kana acquisition.1 At that pace, hiragana fits a one- to two-week block and katakana fits the next. This is also how Genki paces its first two lessons.1

The Three (and a Half) Japanese Scripts

Hiragana: the grammatical glue

Hiragana is a moraic script: each base symbol represents one mora, the timing unit of Japanese, not a phonological syllable.7 The modern inventory has 46 base characters in everyday use; with dakuten (゛) and handakuten (゜) it gains 25 voiced and semi-voiced sounds, and with yōon (small ゃ ゅ ょ) it forms palatalized combinations such as きゃ kya.4

Its main job is grammar. Hiragana writes okurigana (the inflectional endings attached to a kanji stem), grammatical particles, native Japanese words lacking a standard kanji form, and furigana over kanji.4

わたし学生がくせいです。1
"I am a student."

Particles は (topic, pronounced wa), が (nominative), を (accusative, pronounced o), に (dative/locative), and へ (allative, pronounced e) are always written in hiragana regardless of register. Their spelling is fixed by post-war kana usage rules.2

ほんみます。1
"I read a book."

Katakana: the loanword and emphasis script

Katakana shares hiragana's mora inventory; the two scripts encode the same set of sounds with different visual shapes.4 Its standard uses in modern Japanese are gairaigo (foreign loanwords), onomatopoeia, scientific and technical terms (especially species names in biology), and emphasis that functions like italics in English.4

Long vowels in katakana are written with the chōonpu (ー), not by vowel doubling, which is the orthographic difference learners notice first.4

コーヒーをみます。1
"I drink coffee."

Katakana also carries a small set of extended combinations (ティ ti, ファ fa, ヴァ va) used to transcribe non-native sounds in loanwords; these have no hiragana counterpart in standard orthography because they do not occur in native vocabulary.4

いぬがワンワンとく。4
"The dog goes wan-wan."

Kanji: the lexical content

The current 常用漢字表 (jōyō kanji list), issued by the Agency for Cultural Affairs on 30 November 2010, contains 2,136 characters. It functions as a literacy baseline for post-compulsory education and as the permitted character set for government documents.8 The 学年別漢字配当表 (kyōiku kanji list), 1,026 characters since 2020, is the subset taught in Japanese elementary school across grades 1 to 6.5

A single kanji typically carries multiple readings: one or more on'yomi (音読み, readings imported with the character from Chinese, often used in compounds) and one or more kun'yomi (訓読み, native Japanese readings, often used with okurigana for standalone words and inflected forms).4 Because the same character has different readings in different word contexts, kanji are not pronounceable in isolation the way kana are. They are learned through the words that use them.2

べる。1
"Eat."

In 食べる the character 食 takes the kun'yomi ta- and the hiragana okurigana べる carries the verb ending. The same character in a Sino-Japanese compound takes its on'yomi.

食事しょくじをします。4
"I have a meal."

A useful heuristic, with many exceptions: a kanji standing as a word with okurigana usually takes its kun'yomi; the same kanji inside a multi-kanji compound usually takes an on'yomi.49

Rōmaji: a pronunciation aid, not a script

Rōmaji is a romanization scheme, most commonly Modified Hepburn in foreign-aimed materials.2 It is not part of the Japanese writing system in the sense that hiragana, katakana, and kanji are. Full-text Japanese books, newspapers, and websites are not written in rōmaji.2

Inside Japanese IMEs (input method editors), the user types rōmaji. The software converts the keystrokes first into kana and then offers kanji candidates. The displayed and stored text is kana and kanji; the rōmaji exists only as keystrokes.3

Japanese-internal uses of Latin script on signage and in print are limited to station names, place names, brand names, and acronyms (NHK, JR, ANA).2

Why Hiragana Comes First

It encodes Japanese phonology completely

Modern Japanese has five vowels /a, i, u, e, o/, each with a phonemic long counterpart, and roughly twelve to fifteen consonant phonemes depending on the analysis.1011 Every native phoneme can be represented with the kana inventory plus its diacritic modifications.10

The hiragana inventory (46 base, plus dakuten/handakuten, plus yōon) is moraic: each unit is one beat of timing. Reading hiragana aloud therefore trains the correct rhythm of Japanese, which rōmaji does not enforce.7 Long vowels and the geminate small っ (sokuon) each count as a separate mora, and hiragana writes them explicitly.7

大阪おおさかきます。7
"I go to Osaka."

おおさか is four morae: o-o-sa-ka. The doubled vowel is two beats, not one long sound.

一杯いっぱいください。7
"One cup, please."

いっぱい is also four morae: i-(small-tsu)-pa-i. The small っ holds a full beat of silence before the geminated p.

It carries the grammar

Grammatical particles (は wa, が ga, を o, に ni, へ e, で de, と to, から kara, まで made, の no, も mo) are written in hiragana and are obligatory in well-formed Japanese sentences. In writing, they cannot be omitted the way articles can in some other languages.2

Verb and i-adjective inflection is realized through okurigana, which is hiragana attached to a kanji stem. The kanji stem is stable while the kana ending changes across tense, aspect, politeness, and polarity.4 Without hiragana, conjugation is invisible.

FormSpellingMeaning
Plain present食べる"eat"
Plain past食べた"ate"
Plain negative食べない"do not eat"
Polite present食べます"eat (polite)"

The kanji 食 is constant across all four; the hiragana ending alone decides the form.14

It is the bridge to everything else

Furigana, the small hiragana printed above or beside kanji to give their reading, is the standard reading aid in materials for children, graded readers, manga aimed at younger audiences (where blanket coverage is called 総ルビ sōrubi), and dictionaries.4 Japanese monolingual and bilingual dictionaries index entries by their hiragana reading; without hiragana, a beginner cannot look up an unknown kanji-spelled word.4

Why Katakana Comes Second

Same sounds, lower daily frequency

The katakana inventory matches hiragana mora for mora for native sounds. Learning katakana after hiragana is therefore a visual re-mapping task, not a re-learning of the phoneme inventory.4 The second pass goes faster than the first because the sound system is already in place.

High-utility from day one despite the gap

Katakana appears as soon as a beginner steps outside a textbook: brand names, menus, station-area signage, country names, and a substantial share of modern vocabulary borrowed from English, German, French, and Portuguese.4 Delaying katakana much past the first month creates a reading bottleneck against everyday written input.

コンビニでコーヒーをいます。1
"I buy coffee at the convenience store."

アメリカからました。1
"I came from America."

For systematic figures on katakana-vs-hiragana frequency across genres, NINJAL's Balanced Corpus of Contemporary Written Japanese (BCCWJ) is the appropriate primary reference.12

Why simultaneous kana is not the default recommendation

The common pacing convention is to learn hiragana to a confident reading level first, then move to katakana, rather than drilling both kana sets in parallel. This is the order Genki and most mainstream beginner courses present.1

The two scripts share an inventory of sounds, but they also contain a small set of visually confusable shapes inside katakana itself: シ versus ツ and ソ versus ン.413 A learner will meet this confusion set regardless of order. Sequential study lets you consolidate one kana inventory before adding the second visual layer on top.

Why this is a convention, not a finding

There is no peer-reviewed study cited here showing that simultaneous kana acquisition produces lasting confusion. The sequential ordering is a long-standing pedagogical convention, supported by the visual-similarity facts in 4 and 13 but not derived from them. A learner with strong reasons to learn both at once will not necessarily fail.

Why Kanji Comes Third, and Always with Vocabulary

The case for kanji-third

A kanji's readings (on'yomi and kun'yomi) are themselves written in kana when given as headword pronunciations in any standard dictionary or textbook. Learning readings before kana is therefore mechanically impossible at the textbook level.4 Most kanji have two or more readings selected by word context. Learning a character's "meaning" without the vocabulary it appears in therefore does not let a learner produce or recognize that character in connected text.4

The same character behaves differently in different words:

やまのぼります。1
"I climb the mountain."

富士山ふじさんたかいです。4
"Mount Fuji is tall."

The kanji 山 is read yama standing alone (kun'yomi) and -san in the compound 富士山 (on'yomi). The reading is selected by the word, not by the character in isolation.

Why "kanji first" arguments collapse

The strongest pitch for learning kanji before kana is James W. Heisig's Remembering the Kanji, which teaches 2,200 kanji using primitives (decomposed character components) and self-invented mnemonic stories that connect each character's written form to a single English keyword. Detailed stories are given for the first 547 characters. Later characters are given keywords only.14

Any rebuttal has to start with what Heisig's own introduction says the method covers and does not cover:

  • Remembering the Kanji 1 deliberately omits Japanese readings. Heisig's stated premise is to learn writing and meaning first, deferring readings to Volume 2.1415
  • Remembering the Kanji 2 introduces the on'yomi and kun'yomi, organized by reading pattern rather than by character; the imaginative-memory technique of Volume 1 is dropped.15
  • The keywords in Volume 1 are not all conventional glosses. Several are coined for mnemonic purposes, and the book does not consistently mark which keywords are standard and which are invented. Established kanji references such as Hadamitzky and Spahn give the conventional glosses against which Heisig's keywords can be checked.4
What Heisig's book is, and is not

By design, Remembering the Kanji 1 does not teach readings, vocabulary, or grammar; the introduction is explicit about this scope.14 A learner who finishes Volume 1 has 2,200 character forms paired with English keywords, not 2,200 readable Japanese words. The grammar particles, verb endings, and pronunciations needed for connected Japanese are all still ahead.

Learn kanji as part of vocabulary, not as a separate course

The Japanese national curriculum teaches kanji to native-speaker children inside vocabulary lessons, at a rate of 80 in Grade 1 rising to a cumulative 1,026 by the end of Grade 6, with each character introduced through the words that use it.5 Mainstream foreign-learner textbooks follow the same pattern at adult pacing. Genki presents kanji in the vocabulary list of each lesson with the lesson-specific reading printed, not as a parallel character-only track.1

わたし日本語にほんご勉強べんきょうしています。1
"I am studying Japanese."

In 日本語 the characters 日 and 本 each have multiple readings elsewhere, but in this compound they are fixed by the word. That fixed reading, repeated across many words, gradually builds a reader's intuition for which reading to pick.

When (and How) to Switch Off Rōmaji

Rōmaji's legitimate one job

For absolute beginners, Modified Hepburn romanization can serve as a temporary sound-to-symbol bridge while the learner maps the five vowels and the consonant-vowel mora structure onto kana shapes. Genki's first lesson uses rōmaji glosses beneath kana for exactly this purpose, then withdraws them.1

As an input system, rōmaji has a permanent legitimate role: typing rōmaji into an IME to produce kana and kanji.3 The pedagogical question of when to stop reading rōmaji is separate from the IME question.

The hard cutoff: by the end of Genki Lesson 2

Genki I stops printing rōmaji glosses in the main grammar-and-conversation chapters from Lesson 3 onward. The working assumption is that the learner can read hiragana and katakana by then.1 In calendar terms, at the daily cadence the book assumes, that is roughly the end of the second study block. But the textbook-anchored cutoff is the lesson, not the week.

Once a learner can read all 46 hiragana without consulting a chart, every new Japanese input they study should be in kana. Textbook chapters past the introduction assume this.1

What goes wrong if you stay on rōmaji

Rōmaji collapses distinctions that kana preserve. The moraic small っ (geminate), the long-vowel mark and long-vowel kana, and the moraic ん all encode timing that the Latin alphabet does not natively represent.107 Learners reading rōmaji with English phonotactics commonly under-time these morae.

学校がっこうきました。1
"I went to school."

がっこう is four morae: ga-(small-tsu)-ko-o. Written as "gakkou" or "gakkō," English readers tend to compress one beat or the other. The kana representation does not let you.

切手きっていました。1
"I bought a stamp."

The minimal pair きって (kitte, "stamp") and きて (kite, "coming") differs by one mora, the sokuon っ. In rōmaji "kitte" and "kite" look similar to an English-reader rhythm; in kana the timing is unambiguous.

Multiple rōmaji systems coexist (Hepburn, Kunrei-shiki, Nihon-shiki), and they disagree on し (shi versus si), ち (chi versus ti), and つ (tsu versus tu).162 A learner who studies in mixed-rōmaji sources collects inconsistent spellings for the same sound. The kana writings are invariant.

Rōmaji hides moraic timing

The single biggest reason to leave rōmaji behind on schedule is that the small っ, long vowels, and moraic ん each carry a beat of timing that kana mark and rōmaji do not. Months of rōmaji reading installs the wrong rhythm. Re-installing the right one later is harder than learning it from kana on day one.107

A Realistic First-Month Timeline

This is the textbook-aligned shape of the first month, mapped to Genki I's pacing.

BlockFocusDaily time
Week 1Hiragana: 46 base, then dakuten/handakuten, then yōon15–30 minutes
Week 2Hiragana consolidation; rōmaji removed from notes and decks15–30 minutes
Weeks 3–4Katakana on the same drill pattern; first kanji in vocabulary15–30 minutes

Week 1: hiragana, 15 to 30 minutes daily

Genki I, Lesson 1's Reading and Writing section presents the 46 base hiragana, then dakuten/handakuten, then yōon.1 Both stroke-order writing practice and reading practice on kana-only sentences are part of the Lesson 1 workbook material.1

Week 2: hiragana consolidation, rōmaji removed

By the end of Genki Lesson 2, the textbook drops rōmaji glosses from the main Japanese text.1 In practice, the second study block is where a learner removes rōmaji from flashcards, notes, and any reading input that was using it as a crutch.

Weeks 3 to 4: katakana, then early kanji-with-vocabulary

Lesson 2 of Genki I presents katakana on the same drill pattern as hiragana in Lesson 1.1 From Lesson 3, Genki begins introducing kanji in vocabulary lists, with the kanji's lesson-specific reading printed. This is the textbook-aligned first kanji-with-vocabulary block.1

Good to know

Mis-aligning hiragana morae with English syllables

A common early error is reading 学校 gakkō as if it were a two-syllable English word, "GACK-koh," shortening the long vowel and clipping the geminate. The kana writing がっこう records four morae: ga-(small-tsu)-ko-o, with the small っ and the long vowel each given their own full beat of duration.107

The correct form, with both timings explicit:

学校がっこうきます。1
"I go to school."

Japanese is a mora-timed language. Reading kana with English stress-timed rhythm is the single most common pronunciation error in the first few months.

Telling シ from ツ and ソ from ン

The four katakana シ, ツ, ソ, and ン are the universal beginner confusion set. Distinguishing them by overall static shape is unreliable. In handwriting and many fonts, the shapes converge.

The stable cue is stroke direction. シ and ン end with a long stroke written upward, with a low-to-high diagonal motion. ツ and ソ end with a long stroke written downward, top-to-bottom. The short strokes that precede the long stroke lean horizontal on シ and ン and vertical on ツ and ソ. The disambiguation cue is the writing motion, not the resulting picture.413

Furigana is a training-wheels phase, not a final destination

Furigana, the small hiragana printed beside or above kanji, is the standard reading aid in children's books, graded readers, manga for younger audiences, and dictionaries.4 It is a scaffolding tool: useful early, then gradually unneeded as a learner's kanji recognition grows. Adult Japanese readers do not rely on furigana except for rare names and unusual readings; aiming to read without it is the point of learning kanji at all.

Why Japanese keyboards type in rōmaji but display in kana

When you type haru on a Japanese IME and the screen shows はる, you are using rōmaji as a keystroke encoding for kana, not as a writing system. The stored and rendered text is kana; the rōmaji exists only as the keys you pressed.3

The first generation of personal-computer Japanese input used dedicated kana keys on specialized keyboards. Rōmaji input became dominant because it works on a standard QWERTY layout. It did not become dominant because Japanese is "really" written in rōmaji underneath. "I type Japanese in rōmaji every day" is a separate skill from "I read Japanese in rōmaji."

The 2,136 jōyō kanji number is a literacy benchmark, not a finish line

The 2,136 jōyō kanji were defined by the Agency for Cultural Affairs (文化庁) in 2010 as a literacy baseline for post-compulsory education and the permitted-character set for official documents, not as a vocabulary target for Japanese-as-a-foreign-language learners.8 JLPT levels do not correspond one-to-one with the jōyō list. The 1,026-character kyōiku list (MEXT) is the elementary-school subset, taught across six grades alongside vocabulary.5

See also

References

Footnotes

  1. Banno, Eri, Yoko Ikeda, Yutaka Ohno, Chikako Shinagawa, and Kyoko Tokashiki. Genki I: An Integrated Course in Elementary Japanese. The Japan Times, 3rd edition 2020, Lessons 1–2 ("Hiragana", "Katakana"). 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

  2. Backhouse, A. E. The Japanese Language: An Introduction. Oxford University Press, 1993, chapter on the writing system. 2 3 4 5 6 7 8 9 10 11

  3. Lunde, Ken. CJKV Information Processing. O'Reilly, 2nd edition 2009, chapter on Japanese input methods. 2 3 4 5

  4. Hadamitzky, Wolfgang, and Mark Spahn. Kanji and Kana: A Handbook of the Japanese Writing System. Tuttle, revised edition 2012. 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

  5. 文部科学省 (MEXT). 『学年別漢字配当表』(Kyōiku kanji list). Annex to the elementary school curriculum guidelines (revised 2017, effective 2020). https://www.mext.go.jp/ 2 3 4

  6. United States Foreign Service Institute, School of Language Studies. "Foreign Language Training": Language Categories and expected weeks to S-3/R-3 proficiency. U.S. Department of State. https://www.state.gov/foreign-language-training/

  7. Kubozono, Haruo. "Mora and Syllable." In The Handbook of Japanese Linguistics, edited by Natsuko Tsujimura, Blackwell, 1999, pp. 31–61. 2 3 4 5 6 7 8

  8. 文化庁 (Agency for Cultural Affairs). 『常用漢字表』(Jōyō kanji hyō). Official notification published 30 November 2010. https://www.bunka.go.jp/kokugo_nihongo/sisaku/joho/joho/kijun/naikaku/kanji/ 2

  9. Tofugu. "An Introduction to On'yomi and Kun'yomi." https://www.tofugu.com/japanese/onyomi-kunyomi/ (limitation): used only for the on'yomi-in-compounds / kun'yomi-with-okurigana rule of thumb, which is also stated in 4.

  10. Vance, Timothy J. The Sounds of Japanese. Cambridge University Press, 2008. 2 3 4 5

  11. Iwasaki, Shoichi. Japanese (London Oriental and African Language Library). John Benjamins, revised edition 2013, chapter on phonology.

  12. 国立国語研究所 (NINJAL). 『現代日本語書き言葉均衡コーパス』(BCCWJ: Balanced Corpus of Contemporary Written Japanese). https://clrd.ninjal.ac.jp/bccwj/

  13. Tofugu. "How to Tell Apart シ, ツ, ソ, and ン." https://www.tofugu.com/japanese/how-to-tell-apart-shi-tsu-so-n/ (limitation): used for the specific stroke-direction disambiguation cue presented to beginners; the underlying stroke order is confirmed by 4. 2 3

  14. Heisig, James W. Remembering the Kanji 1: A Complete Course on How Not to Forget the Meaning and Writing of Japanese Characters. University of Hawai'i Press, 6th edition 2011, Introduction. 2 3

  15. Heisig, James W. Remembering the Kanji 2: A Systematic Guide to Reading Japanese Characters. University of Hawai'i Press, 4th edition 2012, Introduction. 2

  16. Shibatani, Masayoshi. The Languages of Japan. Cambridge University Press, 1990.