Published August 13, 2026

How to learn katakana, chart first and then the tricky parts

To learn katakana, work through the 46 base characters in gojūon order a row at a time, then immediately attach each one to a real word rather than drilling it alone. Katakana’s own advantage makes this easy: most katakana words are borrowed from other languages, so ア (a) can be learned inside アイス (aisu, ice cream) and エ (e) inside エアコン (eakon, air conditioner). A week or two of short daily sessions covers the chart. What takes longer is recognition speed, because beginner Japanese gives you far less katakana practice than hiragana, so the characters need deliberate review to stop feeling like decoding.

If you have not done hiragana yet, do that first. Hiragana carries the grammar, so almost no Japanese sentence is readable without it. Katakana is the second step, and it is the one that suddenly makes menus, signs, and shop shelves legible.

The katakana chart: all 46 base characters

Read it the way Japanese sorts it, down the vowel column a-i-u-e-o, then across the consonant rows k, s, t, n, h, m, y, r, w. Note the irregular readings: シ is shi not si, チ is chi, ツ is tsu, フ is fu.

aiueo
ア aイ iウ uエ eオ o
kカ kaキ kiク kuケ keコ ko
sサ saシ shiス suセ seソ so
tタ taチ chiツ tsuテ teト to
nナ naニ niヌ nuネ neノ no
hハ haヒ hiフ fuヘ heホ ho
mマ maミ miム muメ meモ mo
yヤ yaユ yuヨ yo
rラ raリ riル ruレ reロ ro
wワ waヲ wo
ン n

ヲ (wo) is in the chart for completeness but you will almost never see it in the wild, since the particle it writes is written を in hiragana. Learn it last and do not worry about it.

Add the voiced marks and the chart grows without adding shapes: two ticks turn カ (ka) into ガ (ga) and ハ (ha) into バ (ba), and a small circle turns ハ into パ (pa). Small ヤ, ユ, ヨ ride along after an i-row character to make one beat: キ (ki) plus small ヤ gives キャ (kya). Those extras are why the full working set is 104 characters rather than 46.

Why katakana feels harder than hiragana

Katakana is not more complicated than hiragana. The shapes are actually simpler, because katakana was built in the ninth century from fragments of kanji, straight strokes clipped out of larger characters, while hiragana came from whole cursive forms. Fewer curves, fewer strokes.

The difficulty is exposure. Hiragana is in every sentence you read, so it drills itself. Katakana shows up only where its jobs are: loanwords, onomatopoeia, species and technical names, and the occasional word set in katakana the way English uses italics. A beginner reading graded material can go a whole page without meeting it. So the characters fade, and the second problem lands on top: several katakana look alike, and unlike hiragana they do not give you much shape to hold on to.

The fix is not more memorising. It is spacing the review out over weeks so that recognition survives the gaps in your reading, which is exactly what a scheduler is for.

Confusable katakana: シ vs ツ and ソ vs ン

Four characters cause most of the trouble, and they are really one problem twice. All four are short strokes plus one long stroke, and the difference is the angle of the short strokes and the direction the long one travels.

シ (shi) and ツ (tsu). On シ the two short strokes sit low on the left, roughly horizontal, and the long stroke sweeps upward from the bottom left. On ツ the two short strokes sit along the top, near vertical, and the long stroke comes down from the top right. Horizontal and rising is shi; vertical and falling is tsu.

ソ (so) and ン (n). Same rule with one short stroke instead of two. ソ has its tick near vertical at the top and the long stroke falls. ン has a flatter, lower tick and the long stroke rises. Vertical and falling is so; flat and rising is n.

Learn these by writing them and the confusion mostly never forms, because stroke direction is the difference and writing is the only practice that has direction in it. Two more pairs worth a second look: ク (ku) against ワ (wa), and ン against ソ inside real words, where a wrong guess turns コンビニ (konbini, convenience store) into nonsense.

The loanword superpower: katakana words you already know

Here is the payoff that hiragana does not offer. Katakana is the script for borrowed words, and modern Japanese has a lot of them. One commonly cited breakdown puts today’s Japanese vocabulary at roughly 33% native words, 49% Sino-Japanese words, and 18% loanwords from other languages, the overwhelming majority of which are written in katakana. Learning the script hands you hundreds of words you effectively already have.

Sound them out and you will hear them: コーヒー (kōhī, coffee), ホテル (hoteru, hotel), テレビ (terebi, television), レストラン (resutoran, restaurant), オレンジ (orenji, orange), チョコレート (chokorēto, chocolate), ニュース (nyūsu, news), ラーメン (rāmen, ramen).

Two honest warnings. First, many are clipped: エアコン (eakon) is “air conditioner” cut in half, コンビニ (konbini) is “convenience store”, リモコン (rimokon) is “remote control”. Expect the front of the word and let the rest go. Second, some have drifted. マンション (manshon) is an ordinary apartment building, not a mansion. ワイシャツ (waishatsu) is a dress shirt, from “white shirt”. Reading the sounds gets you to the door; the meaning still needs checking, which is why katakana words belong in your vocabulary study and not just your reading practice. For where to go after the scripts, see the most common Japanese words.

The long vowel mark and small characters

Three marks do a lot of work in katakana, and all three change how a word sounds rather than what letters it uses.

The long vowel mark ー stretches the vowel before it for one extra beat. ケーキ (kēki) is cake, not “keki”. コーヒー has two of them. This is katakana-specific: hiragana writes long vowels by repeating the vowel character instead.

Small ッ doubles the consonant that follows and takes up a full beat of silence. ベッド (beddo) is bed, ニット (nitto) is knit, リュック (ryukku) is a backpack. Say the pause and the word lands; skip it and it will not.

Small ャ, ュ, ョ glue onto an i-row character to make one syllable, not two. ニュース is nyū-su, three beats, not ni-yu-u-su. And because Japanese has no native v, katakana borrowed ヴ for it, though most words spell the sound with a b-row character anyway.

How Moji - Learn Japanese Daily teaches katakana

Moji for iPhone treats katakana as vocabulary rather than as a separate alphabet drill. All 104 katakana characters are cards in the same study path as everything else, and every card carries a real word to hang the character on: ア with アイス, エ with エアコン, オ with オレンジ, ツ with ツアー (tsuā, tour). You learn the character and leave with a word, which is the difference between a chart you have looked at and a script you can read.

Writing practice for kana is free, with stroke feedback drawn from KanjiVG stroke data, so the シ and ツ problem gets solved the way it should be solved, by drawing them in the right direction rather than staring at them.

Version 1.9.1 added a free Katakana quiz deck alongside a free Hiragana one. The katakana deck deals only katakana, so a learner still working through hiragana is never ambushed by a script they have not started, and it quizzes the reading in romaji, since a syllabary character has no meaning to quiz. It sits in the quiz picker and in the home-screen Quiz widget, which is the practical version of “review it more often than your reading does”.

Underneath, the cards are scheduled by FSRS-6 at a 90% target retention, which matters more for katakana than for most things you will study, precisely because your reading will not refresh these characters on its own. Moji is not a dictionary app. It is a study system with one curated order, and the kana are its first stop.

Katakana is the smallest, fastest win available in Japanese, and it is entirely in Moji’s free tier: 104 character cards with real words, kana writing with stroke feedback, and the Katakana quiz deck. Start with a row a day and see how much of a convenience store shelf you can read by the end of the week: Moji is on the App Store.

Sources