Lesson 5 · The marked kana

Two strokes and a circle

Twenty-five more characters, and not one new shape among them. Every character in this lesson is a character you already write, wearing a mark in its top right corner — and the mark always does the same thing.

≈14 minutes · stroke data from KanjiVG · every word checked against Jisho.org

Lesson 2 ended on a character it could not explain. Everything in its closing table lit up except が — か with two small marks. This is that lesson.

1.Two strokes and a circle

There are exactly two marks, they go in the same corner, and between them they account for all twenty-five characters.

Tenten means “dot dot” and maru means “circle”. Those are the names people actually use; だくてん and はんだくてん are what the marks are called in a textbook.

The tenten is not decoration and it is not arbitrary. It marks voicing — whether your vocal cords are running while the consonant is made. Say ka and then ga with a finger on your throat: the tongue, lips and jaw do exactly the same thing both times. Only the buzz differs. That single distinction covers four whole rows:

Three of those four are the same pairing English makes — k/g, s/z, t/d are voiceless and voiced twins there too. Only h → b looks odd, and that is a fossil: the は row was pronounced with a p centuries ago, so its voiced partner is b and its p sound had to be given a mark of its own. That mark is the maru, and the reason only one row has it is that only one row ever needed it.

2.The twenty-five, each beside its original

Every box carries both characters: the one you learned in lessons 1 and 2 on top, quiet and small, and the marked version underneath at reading size. The romaji line spells out the only thing that actually changed.

Tap any character to see it drawn stroke by stroke and hear it. Watch where the mark falls in the order — it is always written last, after the base character is complete. The tenten is two strokes, the maru is one, so a marked character costs exactly one or two strokes more than the character you already write. Nothing else about the hand movement changes.

3.Three places the pattern leaks

The rule in section 1 is real and it covers almost everything. Three things it does not cover are worth knowing now, because two of them will save you effort and the third will confuse you if nobody says it out loud.

し and ち voice to ji, not to zi or di

These two were already irregular before any mark was added — し is shi and ち is chi, not si and ti. Voicing an irregular sound gives an irregular result. じ is ji. So is ぢ. And づ is zu, exactly like ず.

ぢ and づ are all but dead

Because they duplicate じ and ず exactly, modern Japanese picked one of each and dropped the other. Write じ and ず. You will be right almost every time.

How nearly dead: there is no common word in the language containing ぢ, and none at all containing づ in under three characters. The survivors are the few compounds where a ち or つ got voiced by the word in front of it and kept its own spelling — つづく (to continue), こづつみ (parcel), みかづき (crescent moon, literally a three-day moon, from つき). Recognise them; do not go looking for more.

The maru row hardly ever stands on its own

ぱ ぴ ぷ ぺ ぽ are real, common sounds, but they are almost never the first thing in a word and never sit between two plain vowels. They turn up in two places: after ん, and after the small っ — a held beat of silence you have not formally met yet. さんぽ, きっぷ, えんぴつ, てんぷら. That is why section 5's word list has to break its own rule to teach them at all.

4.Tell them apart at speed

Reading が when it is sitting alone in a box, labelled, next to か, is not a skill. Reading it in half a second with no label is. This drill mixes the twenty-five marked characters back in with the twenty unmarked ones they came from and asks for the sound.

The wrong answers are chosen to hurt: given ば you will be offered ha and pa, because the marks are the entire difference and the eye is trained to skip small things in a corner. Answer by tapping or by pressing 1–4.

A miss shows you the character's story and the words already taught that contain it. Forty-five in a pass; the misses come round again at the end.

5.Seventy-three words the marks unlock

Lesson 4 ended with a list of ordinary words it was not allowed to teach — みず, まど, そば, かぎ, ゆび — because each needed a mark. They are the first five words here.

Same rules as lesson 4, with one addition: two kana, at least one of them marked. A word made only of unmarked characters belongs to the previous lesson, so nothing on this page is a repeat — the two lists share not a single word. Everything is checked against JMdict and only goes in if the dictionary tags it common.

Two characters get no card at all: ぢ and づ. There was nothing common enough to give them, which is section 3's point made in the only way that really lands.

Drill them for meaning

The same three directions as lesson 4 — read it, produce it, and straight into Arabic with no English in between. Wrong answers come from within a meaning group, so a question is only answerable if you know the word.

6.What the marks were worth

Seventy-one characters, and only forty-six shapes. The last twenty-five cost two strokes each and one idea — turn the voice on — and in exchange the number of words you can spell stops being a curiosity and starts being a vocabulary. みず is water. You could not write water yesterday.

The chart is now finished in the sense that matters: there is no further set of characters waiting to be learned before ordinary written Japanese becomes readable. What is left is smaller — the little っ that holds a beat, and the little ゃ ゅ ょ that fuse a syllable — and both are rearrangements of characters already on this page rather than new ones.