Hiragana and Katakana: A Beginner Guide (2026)

🗣️ Kana is the fastest two weeks of study in Japanese — and reading it out loud is where your pronunciation actually begins. SoraSpeak gives beginners unscripted spoken practice with AI Japanese teachers, plus instant corrections on grammar, particles and politeness explained in simple English. Free daily tier, in your browser. Start a free speaking session →

Here's the short version: hiragana and katakana are two alphabets for the same 46 sounds — hiragana carries native words and all the grammar, katakana carries loanwords and sound effects — and most people read both after one to two weeks at 15 minutes a day. Kana is not the hard part of Japanese. It is a small, closed, finite system that you finish and never revisit, which makes it the highest-return fortnight in the entire language.

In this guide: what each script does · why kana beats romaji · the look-alikes · dakuten and small kana · a two-week plan · writing vs recognising · the speaking payoff · the bottom line

What hiragana katakana and kanji each do

Japanese uses three scripts in the same sentence, and they are not competing systems — they divide the labour.

One sentence shows all three at once:

私はコーヒーを飲みます (watashi wa kōhī o nomimasu) — "I drink coffee."

私 and 飲 are kanji doing the meaning. コーヒー is katakana because coffee arrived from outside. は, を and みます are hiragana: the topic marker, the object marker, and the polite verb ending. Strip the hiragana out and the sentence stops being a sentence — that is why hiragana is the script you learn first, and why it is genuinely non-optional even for a speaking-only learner. (It is also why you can get impressively far with almost no kanji at all.)

Hiragana is the grammar glue

Everything that makes Japanese work is hiragana. Particles are single hiragana characters. Verb conjugation happens in the hiragana tail after the kanji: 書く (kaku — "write") becomes 書きます, 書きました, 書きたい. Adjective endings are hiragana: たかい, たかかった. Furigana — the small reading hints printed above kanji — are hiragana too, which means kana alone unlocks a surprising amount of real material.

Katakana is harder than it looks

Beginners usually assume katakana will be easy because the words are borrowed. In practice it is the script people stay slow at, for two reasons.

First, the shapes are more angular and more similar to each other than hiragana's. Second, loanwords are reshaped to fit Japanese sound rules, so recognising them takes a mental step: マクドナルド is Makudonarudo, コンビニ is a clipped konbini, アイスクリーム is aisukurīmu. And not all katakana words come from English at all — パン (pan, "bread") is Portuguese, アルバイト (arubaito, "part-time job") is German Arbeit, ズボン (zubon, "trousers") is French. Guessing from English gets you most of the way and occasionally somewhere wrong.

Katakana also does the job italics do in English. A word written in katakana instead of its usual hiragana or kanji is being emphasised, or made to sound clipped, slangy or technical.

Why kana beats romaji from day one

Romaji feels like a shortcut and behaves like a tax you pay later. Three concrete sounds explain why.

ふ is not an English "f". English f puts your top teeth on your bottom lip. Japanese ふ never touches the teeth — you shape your lips as if blowing out a candle and let air pass between them. Written "fu", your mouth defaults to the English sound every time. Written ふ, it does not.

ら り る れ ろ are not English r, and not l either. The tongue taps the ridge behind your top teeth once and leaves — the same flap as the "tt" in American English butter or ladder. Say "butter" fast, isolate that tap, and you have ら. Read "ra" and you will produce an English r, which sounds distinctly foreign.

う is unrounded. English oo in "food" purses the lips forward. Japanese う keeps them relaxed and almost flat. Romaji "u" hands you the wrong lip shape on the first day and every day after it.

Then there is timing, which romaji hides completely. Japanese is mora-timed: every kana gets roughly one equal beat.

• きて (kite) — "come", 2 beats. きって (kitte) — "postage stamp", 3 beats.
• おばさん (obasan) — "aunt", 4 beats. おばあさん (obāsan) — "grandmother", 5 beats.
• とうきょう (Tōkyō) — 4 beats, to-u-kyo-u. Not the two-syllable "TOH-kyo" that English says.

In kana, the extra beat is visibly a character sitting on the page. In romaji it is a doubled letter or a macron you will skim over. Romaji is not even internally consistent: the same mountain is Fujisan in Hepburn romanisation and Huzisan in Kunrei, and し is variously "shi" or "si".

Use romaji the way this article does — as a reading aid printed beside the kana — for a week or two, then drop it. If you want the wider beginner picture of sounds, patterns and particles in order, our Japanese conversation practice guide for beginners covers the full sequence. For gentle real reading with audio, NHK News Web Easy is free, short and furigana-rich.

The kana pairs that look almost identical

Nearly all kana confusion comes from a dozen pairs. Learn them as pairs, side by side, with the distinguishing rule spoken out loud — never one at a time, weeks apart.

Two shortcuts worth banking. A handful of katakana look like their hiragana partner — か/カ, き/キ, せ/セ, も/モ, や/ヤ, り/リ — and へ/ヘ are effectively the same character in both scripts. That is free progress in week two.

The other shortcut is the stroke-direction rule above. シ/ン go up, ツ/ソ come down. If you can say that sentence from memory, four of the six worst katakana are solved permanently.

Dakuten small kana and long vowels

Once you know 46 shapes, the remaining sounds are built from marks and size changes. No new characters.

Dakuten voices the consonant: k becomes g, s becomes z, t becomes d, h becomes b. きん (kin — "gold") versus ぎん (gin — "silver"); てんき (tenki — "weather") versus でんき (denki — "electricity"). Handakuten appears only on the は row and turns h into p: はん (han — "half"), ばん (ban — "turn, number"), パン (pan — "bread"). Same shape, two dots or one circle, three different words.

Small ゃ ゅ ょ fuse with an i-row kana into a single beat. This is the one that costs people real words: びょういん (byōin) is a hospital, びよういん (biyōin) is a beauty salon, and the entire difference is whether the よ is small. Say きょ as one beat and きよ as two, out loud, until they feel different in your mouth.

Small っ is not a sound — it is a held beat, a gap where the next consonant is doubled. きて (kite — "come") versus きって (kitte — "stamp"); がっこう (gakkō — "school"); ちょっと (chotto — "a little"); in katakana, カップ (kappu — "cup"). English speakers skip it almost automatically, so exaggerate it early: stop your voice for a full beat, then release.

Long vowels take two beats. In hiragana they are written by extending: おかあさん (okāsan — "mother"), おにいさん (oniisan — "older brother"), くうき (kūki — "air"), せんせい (sensei — pronounced sen-sē), おとうさん (otōsan — "father"). A few native words use おお: おおきい (ōkii — "big"), とおい (tōi — "far"). Katakana instead uses a straight bar, the chōonpu: コーヒー (kōhī), ラーメン (rāmen), スーパー (sūpā), ケーキ (kēki).

A realistic plan to learn kana in two weeks

Fifteen focused minutes a day, out loud, in this order. The plan below assumes zero prior knowledge.

Use mnemonics, but make them contain the sound. A picture that does not encode the reading is just a picture. Good ones: き looks like a key. め is an eye with a lash (目 means eye). ふ is Mount Fuji. つ is a tsunami wave curling. ぬ is noodles being twirled. り is a river with two banks. Whether you borrow a published mnemonic set or invent your own matters far less than saying the sound aloud every time you recall the picture.

Test recognition by speed, not by score. A flashcard app that says 95% correct while you take four seconds per card is measuring the wrong thing. The real target is reading any kana in under a second, and the real test is reading a paragraph out loud without hesitating.

How long to learn hiragana, honestly: a week to read it, another week for katakana, and roughly a month before it feels effortless in running text. That last stretch happens automatically while you study grammar and speak, so do not sit and wait for it.

Writing versus recognising which one do you need

These are two different skills, and beginners over-invest in the wrong one.

Recognition is what reading needs, and it is what you should measure. Handwriting is production, and for most modern goals it is optional: Japanese input on a phone or laptop converts romaji keystrokes into kana for you (type "ka", get か), so you can message fluently in Japanese without ever forming a character by hand. The JLPT is entirely multiple choice — you can confirm the exam format on the official JLPT site — so there is no handwriting section to prepare for either.

That said, writing each kana five to ten times when you first meet it is genuinely worth the ten minutes. The motor trace strengthens recall, and more usefully it forces you to notice the differences you would otherwise skim: you cannot write シ and ツ without confronting which direction each long stroke travels.

So: handwrite during the fortnight you are learning, then stop and convert that time into timed recognition and reading aloud. If your goal is conversation rather than calligraphy, that trade pays immediately.

From kana to speaking the pronunciation payoff

Here is the part most kana guides skip. Kana is beat notation. One character equals one mora equals one beat — with two documented exceptions, the small ゃゅょ that fuse into the preceding kana, and the small っ that is a beat of silence. Nothing else in Japanese lays the rhythm out this plainly.

Which means reading kana out loud is not a reading exercise. It is rhythm training, and it is the closest thing beginners have to a pronunciation foundation.

Three drills that convert kana into speech:

• Read rows aloud at an even tempo, tapping one beat per kana on the table. Boring, and it installs mora timing faster than any explanation.
• Read words, then sentences, aloud. Silent recognition lets you "know" ふ while still saying "foo". Only your mouth reveals that.
• Read along with audio. NHK News Web Easy gives you short articles with furigana and slow narration, so you can check your reading against a native speaker in the same minute you produce it. Once you have set phrases under your belt, our list of daily Japanese conversation phrases is good read-aloud material too.

What none of that gives you is correction. You can read ふ with your teeth on your lip for six months and nobody will mention it, because it is still perfectly understandable. The same goes for flattened long vowels, skipped small っ, and — once you start making sentences — the particles, which are the single most correction-hungry part of beginner Japanese. Our companion guide on Japanese particles wa ga wo ni de is the next thing to read after kana.

That correction gap is the specific problem SoraSpeak exists for: you speak out loud in the browser, an AI Japanese teacher replies in natural Japanese with subtitles, and corrections on grammar, particles and politeness come back explained in simple English while the sentence is still fresh in your head. The honest framing: it is a new product, it complements eventual conversation with human beings rather than replacing it, and no software fully substitutes for a good teacher. But for daily out-loud practice with immediate feedback, it is lower-friction than the alternatives. If you would rather compare options first, we did that in our best apps to learn Japanese speaking roundup.

The bottom line

Hiragana and katakana are the same 46 sounds wearing two uniforms: hiragana for native words and grammar, katakana for imports, emphasis and noise. Two accent marks, three small kana, one long-vowel bar, and the system is closed. A week each, at 15 minutes a day, and you are done with it for life.

The mistake is not finding kana difficult. The mistake is postponing it behind romaji, which quietly installs English vowels, English stress and English timing into a language that uses none of them. Every week in romaji is a week of habits to undo later.

So learn the shapes, then immediately do the thing the shapes are for: read them out loud, with real words, at an even tempo. The Japan Foundation publishes learner materials organised around actual communicative use, which is the right next step once kana stops slowing you down.

Two weeks. Fifteen minutes a day. Then everything else in Japanese gets easier at once.

🗣️ Kana gives you the beats. Speaking out loud is what turns them into a conversation. Practise speaking Japanese free →

SoraSpeak is an independent Japanese speaking-practice app. Example sentences are written for learners; natural usage can vary by context and region.

FAQ