Why おおきい has two お and とうきょう has a う

The problem with holding a vowel

In English, "oh" and "o" are different lengths and nobody thinks about it. In Japanese, vowel length changes the word. おばさん（おばさん）means aunt. おばあさん（おばあさん）means grandmother. Same consonants, same first vowel, but the second one is held twice as long. If you type it wrong, you have written about a different relative.

So how do you type "held twice as long"? You do not hold a key down. You type the vowel twice, or you type a specific second kana that acts as the length marker. The trouble is that Japanese uses three different systems for this, and they do not look alike.

おう — the most common disguise

Most long お sounds in native Japanese words are written おう, even though you say a long "oo." This is the one that trips people up, because the spelling and the sound do not match.

• とうきょう（とうきょう）— Tokyo. Type toukyou, not tokyo.
• がっこう（がっこう）— school. Type gakkou, not gakko.
• ありがとう（ありがとう）— thank you. Type arigatou.
• おはよう（おはよう）— good morning. Type ohayou.

Notice that ありがとう and おはよう both end in おう. If you are in a chat in your city or anywhere else, typing arigato will still usually convert to ありがとう on a good IME, but typing arigatou is what the spelling actually is. Learn the spelling, not the shortcut.

おお — the honest one

A smaller group of words spells the long お as おお: two お kana in a row. Here the spelling matches the sound.

• おおきい（おおきい）— big. Type ookii.
• とおい（とおい）— far. Type tooi.
• おおぜい（おおぜい）— many people. Type oozei.

There is no rule that tells you which words use おお and which use おう. It is historical, and the only reliable method is to learn the word. おおきい is おお. がっこう is こう. That is just how they are.

えい — the long え in disguise

The long え sound is usually written えい.

• せんせい（せんせい）— teacher. Type sensei.
• けいかく（けいかく）— plan. Type keikaku.
• えいが（えいが）— movie. Type eiga.

A few words write it as ええ instead, like おねえさん（おねえさん）— older sister. Again, no formula. Just the word.

ー — the katakana bar

Katakana handles long vowels differently. It uses a straight bar, ー, called ちょうおんぷ（ちょうおんぷ）— the long vowel mark. When you type katakana words with a long vowel, you type a hyphen - and the IME converts it.

• コーヒー（コーヒー）— coffee. Type ko-hi-.
• ラーメン（ラーメン）— ramen. Type ra-men.
• ケーキ（ケーキ）— cake. Type ke-ki.
• パーティー（パーティー）— party. Type pa-thi-.

This is why katakana loanwords look clean and hiragana words look messy. The bar is a different tool for a different script. Hiragana almost never uses it.

One more thing about typing

When you type toukyou, the IME does not care that the second う is silent-ish. It just sees the kana and converts. Your job is to type the kana that are actually there, not the sound you hear. とうきょう has a う after と and another after きょ. Type both.

If you remember one contrast from today: おおきい is ookii with two o's, and とうきょう is toukyou with ou. They sound similar. They are spelled differently. That is the whole game.

Yesterday's doubled consonant rule from がっこう still applies here: がっこう is gakkou, with the small っ handled by doubling the k. Long vowels and small っ are the two places where romaji typing stops being obvious.

Tomorrow we leave the keyboard behind and move to the twelve-key flick pad on your phone, where the same word ありがとう takes three taps instead of eight keystrokes. The question is which direction your thumb goes, and why こんにちは is faster there than on a laptop.