Why し is typed ‘shi’ and not ‘si’

Yesterday you typed あいうえお and noticed the pattern: the vowel keys do exactly what they say. Type a, get あ. That pattern holds for most of the keyboard. Today it breaks, four times, and those four breaks are worth memorising on purpose.

The k row: no surprises

Add a consonant in front and you get the row.

• か（か）— ka
• き（き）— ki
• く（く）— ku
• け（け）— ke
• こ（こ）— ko

So きみ（きみ）— you, in the casual sense you would use with a friend — is just kimi. And した（した）— below, underneath — is shita, which we will come back to.

The s row: one exception

Four of the five behave.

• さ（さ）— sa
• す（す）— su
• せ（せ）— se
• そ（そ）— so

Then there is し（し）— shi. Not si. Type si on most input methods and you will still get し, because the system is forgiving. But the spelling you should carry in your head is shi, because that is what the kana actually sounds like, and because every textbook, dictionary and IME hint will show it that way.

This is the first real lesson of today: romaji input is a convention, not a code. It approximates sound. Where sound and alphabet disagree, sound wins.

The t row: three exceptions

• た（た）— ta
• て（て）— te
• と（と）— to

Those three are plain. The other two are not.

ち（ち）— chi, not ti. Type ti and you will usually get ち anyway.

つ（つ）— tsu, not tu. This one catches everyone. The sound sits between "ts" and "tsoo"; there is no clean English equivalent. たこ（たこ）— octopus — is tako, all plain. But つくえ（つくえ）— desk — starts with the awkward one.

Why the input method forgives you

Most Japanese IMEs accept several spellings for the same kana. shi and si both produce し. chi and ti both produce ち. tsu and tu both produce つ. This is deliberate: it accommodates people typing on keyboards built for other languages.

The cost is that you can type for months without ever learning which spelling is standard. Then you sit a test, or read a textbook, or watch a friend in your city type at twice your speed, and the gap shows. Learn the standard spelling now and the forgiving shortcuts become a bonus instead of a crutch.

Reading back is the real skill

Typing is only half of it. The other half is seeing した and hearing "shita" without sounding it out letter by letter.

Try these three, out loud, then check the reading.

• きみ — kimi
• した — shita
• たこ — tako

Now a longer one. すし（すし）— sushi. Two s-row kana, one of them the exception. If you typed susi and got すし, you have already met the forgiving IME. If you typed sushi, you are typing the way the language is actually written.

A note on した

した means "below" or "under". It is also the past tense of する（する）— to do — so you will see it constantly. Same spelling, same reading, meaning decided by context. Nothing to do today except notice it exists.

What to drill

Open any text field and type these without looking at a chart:

か き く け こ さ し す せ そ た ち つ て と

Then type the words: きみ, した, たこ, すし, つくえ.

If you can do that in under a minute, you are ready for the next row — because the next row brings the first kana that is not a simple consonant-plus-vowel at all, and it has a rule that trips up almost everybody the first time.

friend, if you are in your country typing on a laptop keyboard, the same keys apply. Nothing changes but the layout around them.