Hiragana vs. Katakana: How to Tell Them Apart and Use Both
Hiragana and katakana divide the work by function, not by sound. Japanese uses two phonetic scripts that cover the same syllable inventory but do different jobs in a sentence.12 Hiragana carries native words and grammar; katakana carries loanwords, foreign names, onomatopoeia, scientific names, and emphasis. Visually, hiragana is curly and katakana is angular, and real Japanese text mixes both with kanji in the same line.
Overview
The difference between hiragana and katakana is the first structural question learners ask once they have seen the charts. The 46 base hiragana and 46 base katakana encode the same set of syllables, so the question is not "which sound does this script make" but "which script does this job."
Same sounds, different jobs
Modern hiragana has 46 base characters covering five vowels, 40 consonant-vowel syllables, and one nasal coda ん.1 Modern katakana likewise has 46 functional base characters covering the same vowels, the same consonant-vowel syllables, and the nasal coda ン.2
Both scripts encode an identical mora inventory: あ and ア both read /a/, か and カ both read /ka/, ん and ン both read /N/.12 Both also take the same diacritics. The dakuten (゛) and handakuten (゜) marks modify voicing identically across hiragana and katakana, producing 25 additional kana per script (か→が, は→ぱ; カ→ガ, ハ→パ).2
The split is conventional, not phonetic. There is no sound a learner can pronounce in one script but not the other.12
When two scripts cover the same sounds, the choice of script becomes a signal in itself. A Japanese reader registers the conventional spelling as neutral and an unexpected script choice as marked, much as an English reader registers italics as a deliberate departure from upright type.2
This article focuses on two questions: which script does which job in modern writing, and how to tell them apart at sight. Stroke-by-stroke chart coverage belongs in the dedicated chart articles, not here.
Why two scripts coexist (the short version)
Both hiragana and katakana descend from man'yōgana, an 8th-century practice of using Chinese characters phonetically to write Japanese.34 From this shared root, the two scripts evolved in different ways. They served different communities for centuries before the modern division of labor was fixed.
Hiragana evolved as a cursive (sōsho) simplification of whole man'yōgana characters. It was used at first in Heian-era private and vernacular writing; the cursive script was historically associated with women and was called onnade, "women's hand."13 Katakana evolved at roughly the same time (9th century, early Heian period) by a different mechanism. Buddhist monks at Nara extracted single components from regular-script kanji to mark up Classical Chinese texts (kanbun) with Japanese-language reading aids. The name reflects the method: kata "partial, fragmented" plus kana.23
Two scripts survived because they served two different user communities: women's vernacular prose on the hiragana side and monastic kanbun annotation on the katakana side.35 Pre-WWII Japanese used katakana plus kanji as the default mix for official documents, laws, military orders, and telegrams, while hiragana was the marked register for informal or literary writing.25
The 16 November 1946 Cabinet Notification Gendai Kanazukai (現代かなづかい, Cabinet Notification No. 33) reformed modern kana orthography to align with modern pronunciation. In the resulting standard, hiragana became the default for native Japanese words and grammar. Katakana was relegated to loanwords, foreign names, scientific names, and emphasis.67 The 28 June 1991 Cabinet Notification Gairaigo no Hyōki (外来語の表記, Cabinet Notification No. 2) is the current government standard for transcribing foreign words and names in katakana. It includes the extended katakana combinations used for non-native sounds.8
Where this article sits in the reading order
This article assumes you have at least scanned a hiragana chart and a katakana chart once. Dedicated articles cover per-script charts, stroke order, dakuten and handakuten, yōon, and within-script lookalikes.12
The page below is the side-by-side comparison: when each script is used (the functional question) and how to tell them apart at a glance (the visual question).
The functional split: when each script is used
The five katakana domains and the two hiragana domains below cover roughly what a beginner encounters in everyday Japanese text. Each subsection names a function and then shows it in context.
Hiragana: native words and grammar
Hiragana writes "miscellaneous other native words for which there are no kanji or whose kanji form is obscure or too formal."1 In beginner materials, native verbs like taberu "to eat" are typically shown as たべる until the learner adds the kanji 食 and the spelling shifts to 食べる.
Hiragana also carries the grammatical glue in a sentence. The case and topic markers (は, が, を, に, で, へ, と, から, まで) are always hiragana.1 So is okurigana, the kana tail on a kanji-rooted verb or adjective that carries inflection: 食べる, 高い, 行きます. Hiragana is "used to write okurigana (kana suffixes following a kanji root, for example to inflect verbs and adjectives)."1
Conjugation endings, copulas (です, ます, 〜ない, 〜た, 〜て), and sentence-final particles (ね, よ, か) are likewise hiragana.1 One more job: hiragana serves as the standard furigana, the "reading aid that shows the pronunciation of kanji characters."1
私は寒いです。9
"I am cold." (Particle は, adjective ending い, copula です are all hiragana.)
高い山に登ります。9
"I climb a tall mountain."
In a 56.6-million-token sample of the 1993 Asahi Shimbun, hiragana accounted for 36.62% of all characters (83 distinct types). Kanji accounted for 41.38% and katakana for 6.38%.10 Remove hiragana from a Japanese sentence, and what remains is a content-word skeleton with no inflection or particles.110
Katakana: loanwords (gairaigo)
This is katakana's headline function. "In modern Japanese, katakana is most often used for transcription of words from foreign languages or loanwords."2 Gairaigo are Japanese words borrowed from foreign (mostly Western) languages, and "gairaigo are usually written in katakana" in contemporary Japanese.11
毎朝コーヒーを飲みます。11
"I drink coffee every morning."
新しいコンピューターを買いました。2
"I bought a new computer."
Non-English layers are large enough that a learner should not assume "katakana = English." Portuguese contact in the 16th–17th centuries gave パン "bread" (Portuguese pão), タバコ "tobacco" (Portuguese tabaco), and カステラ "sponge cake" (Portuguese castela). Dutch contact during the Edo period gave ガラス "glass" (Dutch glas), コーヒー "coffee" (Dutch koffie), and ビール "beer" (Dutch bier). German technical and medical vocabulary entered in the Meiji era as アルバイト "part-time job" (German Arbeit), テーマ "theme" (German Thema), and カルテ "medical chart" (German Karte). French gave アンケート "survey" (French enquête) and ズボン "trousers" (French jupon).1211
パンとビールが好きです。12
"I like bread and beer."
アルバイトをしています。12
"I am working a part-time job."
The 1991 Cabinet Notification Gairaigo no Hyōki is the current government reference for the katakana inventory used in loanword transcription. It includes the extended combinations (ファ, ヴィ, ティ, トゥ) needed for sounds absent from native Japanese.8
Although katakana accounts for only about 6.38% of running text in a 1993 newspaper sample, its share climbs in advertising, packaging, fashion, food, and technology copy, where loanwords cluster.10
Katakana: foreign personal and place names
Katakana transcribes "country and personal names" of non-Japanese origin. The United States is rendered アメリカ (Amerika).2 The 1991 Cabinet Notification Gairaigo no Hyōki governs the katakana inventory used for transcribing foreign place names and personal names alongside loanword vocabulary.8
ジョン・スミスさんはロンドンに住んでいます。8
"Mr. John Smith lives in London."
妹はフランスとイタリアに行きました。8
"My little sister went to France and Italy."
アンナさんはドイツ人です。8
"Ms. Anna is German."
Japanese names of non-Japanese people who have a Japanese-style name (naturalized citizens, mixed-heritage individuals) sometimes appear in kanji or hiragana. The default for any obviously foreign name is katakana.2 Familiar exonyms (アメリカ, イギリス, ドイツ, フランス, イタリア, ロシア, オランダ) are conventionally fixed and not re-transliterated word by word.8
Katakana: onomatopoeia and mimetic words
Katakana is "also used for onomatopoeia, words used to represent sounds," with ピンポン (pinpon, doorbell ding-dong) given as a canonical example.2 Japanese sound-symbolic vocabulary divides into four main categories: giseigo 擬声語 (animate sounds, like a dog's wan-wan), giongo 擬音語 (inanimate sounds, like rain's zā-zā), gitaigo 擬態語 (states and conditions, kira-kira "sparkling"), and gijōgo 擬情語 (psychological states, doki-doki "heart pounding").13
The conventional script split is approximate, not absolute. Sound-imitating words (giseigo, giongo) tend toward katakana for the harder, sharper feel. State-mimetic words (gitaigo, gijōgo) tend toward hiragana for the softer feel, with frequent crossover when the writer wants the opposite tone.14
犬がワンワンと鳴いています。13
"The dog is barking 'woof woof.'"
玄関でピンポンと音がしました。2
"There was a 'ding-dong' sound at the entryway."
星がきらきら光っています。13
"The stars are twinkling."
雨がザーザー降っています。13
"The rain is pouring down."
Manga and advertising copy use katakana heavily for sound effects because the angular shapes read as loud and emphatic in display type.14 The same mimetic root can appear in either script depending on the writer's tone: きらきら (soft, lyrical) versus キラキラ (sharp, branded, often in product names).1314
Katakana: scientific and technical names
"Technical and scientific terms, such as the names of animal and plant species and minerals, are also commonly written in katakana." The biological binomial Homo sapiens is written ヒト rather than the everyday kanji 人.2 The convention applies even when the species is native and has a long-established kanji or hiragana spelling: ネコ (cat, instead of everyday 猫 or ねこ), イヌ (dog, instead of 犬), and バラ (rose, instead of 薔薇).2
Chemistry, mineralogy, and engineering terms borrowed during the Meiji modernization period largely came through German and arrived in katakana: ポリマー (polymer), カリウム (potassium, from German Kalium), and ナトリウム (sodium, from German Natrium).12 The convention signals "scientific register" to a Japanese reader the way a Latin binomial signals it to an English reader.2
ネコは哺乳類です。2
"The cat is a mammal."
このバラはきれいですね。2
"This rose is pretty, isn't it."
カリウムは金属元素です。12
"Potassium is a metallic element."
Outside the scientific register, the same animals revert to their ordinary spellings: 猫 or ねこ in everyday prose, ネコ in biology texts and pet-food labels that want a technical or branded feel.2
Katakana: emphasis (the italics-equivalent)
Japanese has no italics, so katakana often fills that slot. Words the writer wants to emphasize, mark as ironic, or set apart visually take katakana, "mirroring the usage of italics in European languages."2 A native word that would normally be hiragana or kanji may appear in katakana to sound sharp, casual, urgent, advertorial, or "foreign-flavored": ヤバい instead of やばい, オレ instead of 俺, ダメ instead of だめ.2
あれはヤバいよ。2
"That's bad news."
オレが行く。2
"I'll go."
Company and brand names use katakana to stand visually apart and read as a stylized whole rather than as semantic kanji. スズキ, トヨタ, and ソニー coexist with or replace the underlying kanji surnames (鈴木, 豊田) in branded contexts.2
ソニーの新しいテレビです。2
"It's Sony's new television."
For a beginner, this case is a recognition target, not a production rule. Learners who default to the convention (native = hiragana or kanji, loanword = katakana) read as fluent. Reaching for ヤバい or オレ in writing requires a tone judgment that comes with intermediate reading exposure.2
This section is only a preview. The full inventory of stylistic native-word-in-katakana usage lives in its own dedicated article.2
Edge cases worth knowing
A few peripheral uses of katakana shape the overall picture. They also explain why all-katakana text turns up in unexpected places.
Pre-WWII katakana-dominant text. Before the 1946 reform, Japanese laws, military orders, and government documents were typeset in katakana plus kanji. Hiragana was the marked, informal register. Telegrams in Japan were transmitted entirely in katakana until 1988.25
Legacy computer encoding. Early 8-bit Japanese character sets, notably JIS X 0201 issued in 1969, handled only half-width katakana alongside ASCII (the basic Latin-letter computer character set). Older computer interfaces, point-of-sale receipts, and ATM displays therefore often rendered Japanese entirely in katakana. The look survives where legacy systems persist.2
Stylized speech in fiction. Manga, anime, and novels routinely render the speech of robots, foreigners, drunks, and aliens in katakana to mark it as accented, mechanical, or non-native. The same trick appears for dialect renderings the writer wants to flag as outside the standard.2
Brand and trade names. Many Japanese companies write their names in katakana even when a kanji form exists. ソニー is the official corporate logotype. トヨタ appears in marketing alongside the kanji 豊田 used on legal documents.2
The visual difference: how to tell them apart at a glance
The functional split tells you why two scripts exist. This section tells you how to recognize which is which at sight.
Hiragana is curly, katakana is angular
The shape contrast comes from the history described above: hiragana from cursive (sōsho) renderings of whole man'yōgana characters, katakana from clipped components of regular-script kanji.123 The beginner's visual test is short. Round, continuous loops and flowing curves point to hiragana; sharp corners, straight strokes, and clean angles point to katakana.12
There are outliers. A handful of katakana include one curve, and a handful of hiragana include one sharp turn. Still, the curly-versus-angular generalization holds for the great majority of the 92 base characters.12
Side-by-side comparison
The table below maps each row of the gojūon (the standard kana order) across both scripts. The romaji column uses Modified Hepburn.12
| Row | Hiragana | Katakana | Romaji |
|---|---|---|---|
| a | あ い う え お | ア イ ウ エ オ | a i u e o |
| ka | か き く け こ | カ キ ク ケ コ | ka ki ku ke ko |
| sa | さ し す せ そ | サ シ ス セ ソ | sa shi su se so |
| ta | た ち つ て と | タ チ ツ テ ト | ta chi tsu te to |
| na | な に ぬ ね の | ナ ニ ヌ ネ ノ | na ni nu ne no |
| ha | は ひ ふ へ ほ | ハ ヒ フ ヘ ホ | ha hi fu he ho |
| ma | ま み む め も | マ ミ ム メ モ | ma mi mu me mo |
| ya | や ゆ よ | ヤ ユ ヨ | ya yu yo |
| ra | ら り る れ ろ | ラ リ ル レ ロ | ra ri ru re ro |
| wa | わ を | ワ ヲ | wa wo |
| n | ん | ン | n |
Pairs that share a shape concept
A small number of hiragana and katakana pairs come from the same kanji source and keep a family resemblance the learner can use as an anchor.129
The か/カ pair both derive from 加. The hiragana keeps the full cursive sweep; the katakana keeps the left-side "force" component.12 The せ/セ pair both derive from 世. The cross-bar shape is visible in both scripts.12
The へ/ヘ pair both derive from 部 (right-side component 阝→へ). The two are near-identical in modern fonts and distinguishable mainly by context.12 The り/リ pair both derive from 利. Hiragana keeps the connecting stroke; katakana drops it for two clean parallels.12 The や/ヤ pair both derive from 也.12
The remaining pairs come from different source kanji and look visually unrelated. That is why a learner who has memorized hiragana cannot read katakana by inference alone.9
Cross-script lookalikes to watch for
A few hiragana and katakana cross-script pairs look so similar that beginners misread them. The diagnostics below focus on stroke details. A learner who internalizes them rarely confuses the pair again.9
- へ / ヘ. Functionally identical in most digital fonts; they differ only in slight slant or stroke thickness in some handwritten or display styles. Read by context: if the character is acting as a directional particle, it is the hiragana へ.12
- り / リ. Hiragana り is a single connected curve broken into two strokes that lean toward each other. Katakana リ is two clearly separate vertical strokes with no curve.9
- や / ヤ. Hiragana や has a curving stroke that hooks back. Katakana ヤ has a clean diagonal with a short cross-stroke.9
- せ / セ. Hiragana せ has a hooked finish on the bottom-right. Katakana セ ends in a clean diagonal.9
- か / カ. Hiragana か has a curved second stroke and includes a small mark to the upper right of the cross. Katakana カ is just two angular strokes with no mark.9
母へ手紙を書きます。9
"I will write a letter to my mother."
ヘルメットをかぶります。8
"I'll put on a helmet."
The first uses hiragana へ as a directional particle. The second opens with katakana ヘ as the first character of a loanword. The shapes in most fonts are identical; only the role tells them apart.
りんごとリンゴは同じ意味です。2
"りんご and リンゴ have the same meaning."
Same word, two scripts: hiragana for the everyday fruit, katakana for a scientific or stylized rendering.
Reading mixed text: hiragana, katakana, kanji, rōmaji
A real sentence, annotated
Modern Japanese mixes four scripts in a single sentence and uses no spaces. Readers parse by script change as well as by lexical knowledge.10 The sentence 私はコーヒーを飲みます ("I drink coffee") breaks into script-tagged pieces. Each piece does a recognizable job.
私はコーヒーを飲みます。10
"I drink coffee."
Kanji carry the content words (私, 飲), hiragana carries the grammar and verb inflection (は, を, みます), and katakana flags the foreign-origin word (コーヒー, from Dutch koffie).1210
Tシャツを3枚買いました。10
"I bought three T-shirts."
This second sentence stacks four scripts: a Latin letter (T), katakana (シャツ), hiragana (を, いました), and kanji (枚, 買), plus an Arabic numeral (3). Real Japanese mixes them without apology.
What the script switches signal
Japanese is written without spaces between words; "Japanese lacks spaces between words. Script transitions function as implicit boundaries."10 A worked segmentation given on Wikipedia is あなたはお母さんにそっくりね → あなた / は / お母さん / に / そっくり / ね. Each boundary aligns with a script change or a particle-shape change.10
The reader's eye uses the kanji-to-hiragana transition as a content-word-to-grammar cue. It uses a kanji-or-hiragana-to-katakana transition as a "this word is a loan, name, or emphasized form" cue.10 Naming this pattern helps beginners parse unfamiliar sentences faster. Even with unknown kanji in the way, the script boundaries hint at where one word ends and the next begins.10
When reading new Japanese, scan first for the script transitions. The point where kanji gives way to hiragana usually closes a content word. The point where hiragana or kanji gives way to katakana usually opens a loanword or name. Spaces never come; the script change is doing that work.10
When rōmaji shows up
Rōmaji (Latin letters) is the fourth script in modern Japanese. It appears in acronyms (NHK, JR, TV), some company logos, and learner materials. In the 1993 Asahi Shimbun corpus, rōmaji accounted for 0.46% of tokens.10
Rōmaji is not a fifth meaning-bearing system. It is a transliteration of either Japanese (Hepburn, Kunrei, Nihon-shiki) or already-Latin-script foreign material (proper names, technical abbreviations).10 Learner materials use rōmaji as scaffolding for the pre-kana stage. The conventional pedagogical advice is to stop relying on rōmaji once kana are fluent.9
Good to know
The two scripts cover identical sounds, so transliteration goes both ways
The 46 base hiragana and 46 base katakana cover identical mora inventories. In principle, any Japanese word can be written in either script.12 Convention overrides that option. Writing りんご "apple" in katakana (リンゴ) is grammatically fine but reads as scientific or stylized. Writing コーヒー "coffee" in hiragana (こーひー) reads as childish, archaic, or visually broken.2
This is literacy, not preference. A fluent reader registers the unmarked spelling as neutral and the unexpected script choice as marked.2
"All-katakana" Japanese is a real register, not a beginner training mode
Pre-WWII Japanese laws, military communications, and most official documents combined katakana with kanji. Hiragana served the role katakana plays today.25 Telegrams in Japan used katakana-only Japanese until 1988.2 Legacy computer systems handled single-byte half-width katakana but not hiragana, leaving an all-katakana trail in older receipts, ATM screens, and point-of-sale (POS) systems.2
A learner who encounters an all-katakana document is reading authentic historical or technical Japanese, not beginner simplification.2
The katakana-for-emphasis case is real but secondary
Native words written in katakana (ヤバい instead of やばい, オレ instead of 俺, ダメ instead of だめ) signal stylistic emphasis or casual register.2 For the beginner, this is a recognition target, not a production rule. The convention to follow when writing is native = hiragana or kanji, loanword = katakana. Stylistic katakana of native words requires a tone judgment that comes with intermediate reading exposure.2
Loanwords are not always from English
Japanese has layers of loanwords from before the post-WWII English wave. Portuguese contact in the 16th–17th centuries gave パン "bread," タバコ "tobacco," and カステラ "sponge cake." Dutch contact during the Edo period (1603–1853) gave ガラス "glass," コーヒー "coffee," and ビール "beer." German contact in the Meiji period (late 19th–early 20th century) gave アルバイト "part-time job," テーマ "theme," and カルテ "medical chart." French contact from the Meiji period onward gave アンケート "survey" and ズボン "trousers."1211
Recognizing the source language helps predict spellings. An unfamiliar katakana word ending in -ウム is almost certainly a Meiji-era German scientific loan (カリウム, ナトリウム, ラジウム).12 Post-WWII gairaigo are overwhelmingly from English, but treating "katakana = English" as a rule misses about three centuries of older borrowings still in daily use.1211
Mnemonic origins: a few katakana look like their hiragana siblings
Five pairs share a kanji source and keep a visible family resemblance: か/カ (from 加), せ/セ (from 世), へ/ヘ (from 部), り/リ (from 利), や/ヤ (from 也).129 The remaining roughly 41 pairs come from different source kanji and have to be learned independently. Many learners find katakana harder to retain than hiragana partly because these cross-script memory hooks cover only a small fraction of the inventory.9
Within-script confusions are a different problem
Confusing シ with ツ, or ぬ with め, is a within-script lookalike issue, not a hiragana-versus-katakana issue.2 The diagnostics differ from the cross-script case. Within-katakana lookalikes are resolved by stroke direction (シ and ン start bottom-up; ツ and ソ start top-down), while within-hiragana lookalikes are resolved by loop count and crossbar position.9
See also
- Hiragana, Katakana, or Kanji First? A Beginner's Script Order
- Long Vowels in Katakana: How the Chōonpu ー Works and Why Hiragana Doesn't Use It
- Long Vowels in Hiragana: How to Read and Write ああ, いい, うう, ええ/えい, and おう/おお
- Katakana Mnemonics That Actually Work
- Hiragana Mnemonics That Actually Work
- Japanese Punctuation and Typographic Conventions