Hiragana vs Katakana: The Difference, the Look-Alikes, and Which to Learn First

What hiragana is for: the glue in every sentence

Hiragana does the structural work. It writes native Japanese words, the particles that hold a sentence together, the endings on verbs and adjectives, and the reading hints printed beside unfamiliar kanji. Take hiragana out of a Japanese page and the sentences stop working as sentences.

Four jobs, in roughly the order you'll meet them.

Native Japanese words. Words with no kanji, or whose kanji is too obscure or too formal for daily use, get written in hiragana: ここ (koko) here, ありがとう (arigatō) thank you, きれい (kirei) pretty. That last one does have a kanji form, 綺麗, and dictionaries mark it as usually written in kana alone.

Grammar particles. These short words mark who is doing what to whom, and they're written in hiragana every single time.

Two of those shift their reading on the job. は is normally 'ha' but becomes 'wa' as the topic particle, and を is read 'o'. The hiragana chart works through that rule in full.

Verb and adjective endings, called okurigana. Japanese hangs the changeable part of a word off a kanji stem and writes it in hiragana. 食べる (taberu) means to eat: 食 stays as kanji, べる is hiragana. Change the tense and only the hiragana moves, giving 食べます (tabemasu) and 食べた (tabeta). The kanji sits still and keeps carrying the meaning.

Furigana. Small hiragana printed above or beside a kanji to spell out its reading. Children's books, manga and beginner textbooks lean on it constantly, which is one more reason hiragana pays off early.

What katakana is for: the import lane

Katakana handles what arrived from outside, plus a couple of special effects. Its range is narrower than hiragana's, which is exactly why it feels less familiar at first. Five situations cover almost everything you'll see.

1. Loanwords (gairaigo). Words borrowed from other languages, and by a wide margin katakana's biggest job: コーヒー (kōhī) coffee, テレビ (terebi) TV, パン (pan) bread, which came from Portuguese pão rather than English.

2. Foreign names and places. Your own name will be written in katakana, along with every country and city outside Japan: アメリカ (Amerika) America, マイク (Maiku) Mike.

3. Onomatopoeia. Sound effects, which Japanese uses far more heavily than English does, especially in manga: ワンワン (wanwan) a dog's woof, ドキドキ (dokidoki) a pounding heart.

4. Scientific and technical names. Animal and plant species names appear in katakana in biology writing and on packaging, even when the creature is thoroughly Japanese: ネコ (neko) cat, イヌ (inu) dog.

5. Emphasis. Katakana works the way italics do in English. Write a normally-hiragana word in katakana and it pops off an advert or a menu. This is the use that surprises learners, because the word isn't foreign at all, it's just being said a little louder.

The full set of shapes lives on the katakana chart, along with the extra combinations Japanese built to spell sounds it didn't originally have, like ファ (fa) and ティ (ti).

Hiragana vs katakana side by side: the look-alike pairs

The shapes aren't a style choice, and the resemblances between the two scripts aren't random either. Both descend from man'yōgana, Chinese characters that early Japanese writers used purely for their sounds. Hiragana came from writing those characters in the flowing cursive style called sōsho, so a whole kanji got smoothed into one continuous curve. Katakana came from Buddhist monks in the 9th century who needed fast margin notes, so they clipped one component out of a character and used that alone. The name records it: katakana means fragmentary kana.

Compare か and カ. Both are 'ka' and both descend from the kanji 加. Hiragana か is the entire 加 written cursively, which is why it flows and closes up on itself. Katakana カ is only the left-hand part of 加, which is why it stops after two straight strokes.

That's also why certain pairs across the scripts look like cousins: they are. Study them in pairs rather than one at a time, because your eye learns a shape fastest by contrast.

Read the pattern off that last column, because it does real work. Where the ancestor is shared, the two characters are variations on one shape, so one rule separates them: find the curve, and the curved one is hiragana. Where the ancestors differ, as with さ and サ or ち and チ, the resemblance is coincidence, and the two shapes have less in common than they first appear.

へ and ヘ are the extreme case. Same source kanji, same sound, drawn almost identically. Nobody reads those character by character. You read them from the script around them, the way you tell an English 'l' from a '1' by the company it keeps.

A quick sanity check sits underneath all of this. Katakana is built from straight strokes and hard corners, while hiragana rounds and loops. The one katakana that closes up, ロ (ro), does it with four square corners.

The four katakana that trip up everyone: シ, ツ, ソ, ン

This confusion outlasts every other one, and it happens entirely inside katakana. シ (shi), ツ (tsu), ソ (so) and ン (n) are built from the same two ingredients: a few short dashes and one long sweeping stroke. Wikipedia's katakana article puts it plainly, noting that the four look very similar in print apart from the slant and the stroke shape.

Sort them in two steps.

Step one: count the dashes. シ and ツ have two. ン and ソ have one. That splits four characters into two pairs before you've thought about anything else.

Step two: read the direction. This is the step that actually solves it, and it comes straight from how the characters are written by hand.

Say it out loud until it sticks: shi and n go up, tsu and so go down.

The dash angle is the clue you can use on printed text, where there's no hand to watch. シ and ン carry flat, horizontal dashes down the left side of the character. ツ and ソ carry upright, vertical dashes along the top. That difference shows up more clearly in brush writing than in a screen font, which is why the pair feels harder online than on paper.

Then drill it with one word: シャツ (shatsu), shirt. It opens with シ and closes with ツ, so a single three-character word puts both members of the hardest pair right next to each other. Read it correctly ten times and the problem is done.

Hiragana has its own look-alikes, れ against ね against わ, る against ろ, and ぬ against め. They're gentler than this set, because no two of them share a stroke pattern quite so closely. The hiragana chart groups them for study.

Hiragana or katakana first? Hiragana, and it isn't close

Learn hiragana first. The honest reason isn't that katakana is harder, because it isn't. It's that you can read a surprising amount of Japanese while quietly skipping every katakana word, and you can't skip a single hiragana one.

The numbers back this up. A frequency study of a full year of the Asahi Shimbun newspaper, around 56.6 million tokens, found kanji made up 41.38% of characters, hiragana 36.62%, and katakana only 6.38% (Chikamatsu and colleagues, Behavior Research Methods, Instruments, & Computers, 2000). Hiragana turns up close to six times as often as katakana.

Frequency is only half the argument. Hiragana carries the grammar, so a sentence missing its particles and verb endings isn't a sentence with gaps in it, it's unreadable. A sentence missing its katakana words is a sentence with a few unknown nouns, which you can still parse and often guess from context.

Hiragana also unlocks your study materials. Furigana is written in hiragana, dictionaries give readings in hiragana, and beginner textbooks assume it from chapter one. Learn it first and everything else you own starts working.

Two honest caveats. The gap is smaller than the ordering implies, because the scripts share identical sounds: once you read hiragana, katakana is new shapes for sounds you already own, usually days of work rather than weeks. And don't push it back by months. Loanwords are on the first menu and shop sign you meet, so a katakana-blind learner is stuck exactly where a beginner most wants to read.

Hiragana vs katakana vs kanji: where the third script fits

Hiragana and katakana are two thirds of the story. Kanji is the third script, borrowed from Chinese, and it runs on a different principle: each character carries meaning rather than a fixed sound, and most can be read more than one way depending on the word.

The scale is different too. The jōyō list, the characters Japanese schooling and publishing treat as standard, holds 2,136 kanji as of its 2010 revision. Of those, 1,026 (the kyōiku kanji) are spread across the six years of elementary school.

Kanji and hiragana work as a pair, not as rivals. Kanji supplies the stem and the meaning, hiragana supplies the ending that inflects it. Kanji also does a job nothing else can: Japanese leaves no spaces between words, so the visual switch from a dense kanji to a run of hiragana is what shows a reader where one word stops and the next starts. A sentence written entirely in hiragana is readable and genuinely tiring.

Katakana stands outside this arrangement, which restates the split one last time. Hiragana and kanji build the sentence. Katakana labels what came from somewhere else. Wikipedia's overview of the Japanese writing system has the historical detail.

One sentence, three scripts

Here's the whole system doing its job in one short sentence.

私はコーヒーを飲みます。

Watashi wa kōhī o nomimasu. 'I drink coffee.'

Now watch what changes when the sentence does.

Put it in the past and you get 私はコーヒーを飲みました (watashi wa kōhī o nomimashita), 'I drank coffee'. Only the hiragana moved. 飲 didn't budge, because the meaning didn't.

Swap the drink for tea and you get 私はお茶を飲みます (watashi wa ocha o nomimasu). The katakana disappears completely, because tea isn't a loanword in Japanese, so お茶 (ocha) is written with a hiragana prefix and a kanji instead.

That's hiragana vs katakana in one experiment. Hiragana flexes, kanji holds still, and katakana shows up only when the sentence reaches outside Japanese for a word.

A practice idea. Find a Japanese menu, product label or shop sign online and sort every word into three buckets by script. You don't need to know what any of it means. The katakana entries jump out because they're angular, and each one is a foreign word you can probably decode by sounding it out. Then check yourself against the full tables: the hiragana chart for all 46 rounded characters and the katakana chart for all 46 angular ones. Next lesson: Japanese numbers uses both scripts from its first line, and the Japanese hub lays out the rest.