Reference sheet

Katakana stroke chart

All 71 katakana — the 46 basic characters and the 25 that wear a tenten or a maru — every stroke drawn and numbered in writing order, each one beside the hiragana that shares its sound and the kanji it was cut from. Built to print, and meant to live beside the paper.

Stroke data: KanjiVG, © Ulrich Apel, CC BY-SA 3.0 · ancestry cross-read across six sources, English and Japanese (section 6)

1.The grid

The same rows as the hiragana sheet, spelling the same sounds. Read it as a·i·u·e·o and then those five vowels with each consonant in front. The gaps in the y and w rows are real: those syllables either never existed or were abolished in 1946, so the script has forty-six characters rather than fifty. Nothing about the sound system differs between the two scripts — only which words get written in which.

The last five rows are the first rows wearing a mark, not new shapes: two small strokes (the tenten) voice a consonant, and one small circle (the maru) turns h into p. Identical to the hiragana rule — see the tenten and maru sheet. ジ is ji, not zi.

ヂ and ヅ are not used. They duplicate ジ and ズ exactly and the dictionary has no katakana word containing either, so their cards below carry no word list. When you hear ji or zu, write ジ or ズ.

One extra character, off the chart: ヴ — ウ with a tenten, giving a v for foreign words. Hiragana has no equivalent. You will meet the same word spelled both ways (ヴァイオリン and バイオリン are both current), so read it but do not worry about writing it.

Irregular readings — the same three as in hiragana: シ is shi, チ is chi, ツ is tsu. Everything else is exactly consonant + vowel.

ヲ is on the chart but not in use. Modern Japanese writes the object particle を in hiragana and spells a foreign wo as ウォ, so ヲ has no word list below and no entry in the dictionary. Learn it to read pre-war text; do not look for a use for it.

Not on this sheet: the long-vowel mark ー, which holds the preceding vowel for one more beat (コーヒー is four beats, not two) and appears constantly in real katakana; the small kana; and the tenten and maru, which katakana takes exactly as hiragana does. The words in section 4 use ー where the spelling requires it, and lesson 9 § 7 explains it.

2.Stroke order

Numbers mark where each stroke begins — which matters more here than in hiragana, because シ against ツ is decided by direction alone. The same three habits apply: top before bottom, left before right, horizontal before the vertical that crosses it.

Under each character: its romaji, its hiragana twin, the stroke count, and the man'yōgana kanji it was cut from. One kanji, in colour, means the pair are siblings — both scripts came from that same character, so the resemblance between them is real. Two kanji means they are unrelated and any likeness is coincidence. A question mark means the ancestry is unsettled, and the card names the candidates rather than picking one.

The hiragana equivalents are on their own sheet: hiragana stroke chart.

3.How the two scripts were made

Both were built in the Heian period from man'yōgana — Chinese characters used for their sound rather than their meaning. The difference is what was done to them, and it accounts for every shape on this sheet.

Twenty-nine of the forty-six pairs descend from one shared ancestor, and there the family likeness is inherited and can be trusted: ウ オ カ キ ク コ シ セ ソ テ ト ナ ヌ ネ ノ from the first half, and ヒ フ ヘ ホ メ モ ヤ ユ ヨ ラ リ レ ロ ワ from the second. Thirteen came from different kanji entirely and have no reason to look alike. Four are unsettled.

The unsettled four split into two kinds. ツ and ン have no agreed ancestor: 川, 州, 津 and 闘 have been proposed for the first, and 尓, 无 and 二 for the second — where the strongest candidate may not be a kanji at all, but the old annotation mark for a nasal beat. ニ and マ have two candidates each, and in both cases the choice decides whether the pair are siblings at all: 仁 or 二 for ニ (and に is certainly 仁), 末 or 万 for マ (and ま is certainly 末). Both shapes are agreed even though their descent is not.

The opposite caution applies to a handful of the twenty-nine: the likeness is so faithful that it causes its own error. In the writing study behind section 4, the characters children most often replaced with the hiragana of the same sound were モ, ヤ and リ — all three genuine siblings. And ハ, ミ and モ get confused with the kanji 八, 三 and 毛, which is unsurprising given that those are the very characters they were cut from.

4.Confusion pairs to watch

Straight strokes are quick to write and easy to mistake for one another. Writing both members of a pair side by side fixes them faster than looking at them does.

Two characters fail differently. ヨ and ミ are the two most often written mirrored — 18 and 10 children out of 430. Nothing is being confused with anything; the stroke simply starts or attaches on the wrong side. ヨ's horizontals all meet the upright on the right, and all three of ミ's strokes sweep right to left.

And ハ, ワ, ロ, ユ and ヨ (with ヌ and ネ from the first half) lead the study's blank answer column — not confused, simply not recalled. Those need production practice from the sound, not recognition practice from the chart.

5.What katakana is for

Which script a word is written in says nothing about how it sounds. It is a label on the word, and reading it saves you time.

1. Loanwords (gairaigo) — by far the commonest use, along with foreign place names and personal names.
2. Onomatopoeia — words that are sounds.
3. Technical and scientific terms, including the scientific names of animal and plant species, and minerals. This is why ネコ (cat) and スイカ (watermelon) turn up in katakana although they are native words — but as an ordinary word an animal or plant name takes its kanji if the character is on the standard list, so 犬 stays 犬. イヌ is nearer to “Canis familiaris” than to “dog”. The mechanism is contrast, not biology: before the war, everyday text was kanji + katakana and species names were set in hiragana; the post-war reforms flipped both.
4. Company names, distinguishing a firm from the surname it shares.
5. Signs and advertising.
6. Emphasis — the job italics do in English.

Reading trick: because most katakana words are borrowed, saying the syllables aloud and letting your ear finish the word works far more often than it should. テレビ is terebi is television.