Katakana to English: The Loanwords Hiding in Japanese

Japan did not borrow a handful of English words — it borrowed tens of thousands, and wrote every one of them in katakana. Which means a real share of any Japanese page you open is English you already know, spelled out in a script you cannot yet read.

That is the whole trick of this article. コンピューター is konpyuutaa. Say it out loud and it is "computer". テレビ is terebi, "television". Once you can convert katakana to romaji, an unexpectedly large amount of Japanese stops being opaque and turns into English wearing a heavy accent.

The accent is the interesting part. It is not random — it is a small set of rules that English sounds cannot survive, and learning them is what turns "I can almost hear it" into "I can read it".

Why katakana exists for this

Japanese has three scripts, and they divide the labour cleanly. Hiragana writes native Japanese words. Kanji carries meaning. And katakana writes foreign vocabulary — loanwords imported from other languages, overwhelmingly English in the modern era.

So katakana is a signal. When a word appears in katakana and looks meaningless, your first assumption should be that it is not Japanese at all. It is something else, respelled. In older text the source might be Portuguese, Dutch, German or French — アルバイト (arubaito, "part-time job") comes from the German Arbeit, not from English — but in anything written today it is usually English.

The category has a name: 外来語 (gairaigo, "words from outside"). And it is large enough that a beginner who learns to read katakana has effectively unlocked a free vocabulary subsidy.

The four things katakana does to English

Every loanword has been forced through the same filter: Japanese has five vowels and about fourteen consonant sounds, and English has far more of both. The gap is what makes the results unrecognisable. Four rules cover most of it.

Long vowels get a ー. The dash extends the vowel before it, so a long English vowel becomes two beats: コーヒー koohii (coffee), ビール biiru (beer), ケーキ keeki (cake). When you see ー, hold the vowel.

Small ッ doubles the next consonant. ベッド is beddo (bed), カップ is kappu (cup), サッカー is sakkaa (soccer). The small tsu is not a sound of its own — it is a beat of silence that lands on the consonant after it.

There is no L. Japanese has a single liquid consonant that sits between English r and l, so both letters collapse into the same row: ライト raito is either "light" or "right", and only context tells you which. This is why Japanese speakers swap the two in English, and why you cannot decode a loanword by assuming the r is an r.

Consonants cannot end a word. With one exception — ん (n) — every Japanese syllable is a consonant plus a vowel. English words that end in a consonant therefore gain a vowel to carry the final sound: ベッド beddo (bed), ミルク miruku (milk), ベスト besuto (best), タクシー takushii (taxi). The added vowel is usually u or o, and it is very quiet. Say it and drop it, and the English reappears.

Put the rules together and the workflow is: read the katakana as romaji, restore the vowels that English swallowed, and listen. ストロベリー → sutoroberii → "strawberry".

The ones you cannot guess: abbreviations

Japan shortens long loanwords ruthlessly, and the result is often impossible to reverse without knowing the original.

These are not words with a heavy accent. They are English cut down to a stub, and パソコン in particular will defeat you until someone tells you what it stands for. Learn them as vocabulary, not as sound changes.

和製英語: words that look English and are not

Here is the trap that catches people who have got good at this. A large set of katakana words are English-shaped but not English — coined in Japan out of English parts, or borrowed and then given a meaning English never had. The Japanese term is 和製英語 (wasei eigo, "English made in Japan").

Read the middle column again. Every one of these is a real, common, everyday Japanese word. Every one of them is also a word an English speaker would not use, or would understand completely differently — マンション is a perfectly ordinary apartment building, and calling it a mansion will get you a very different picture from the one the speaker has.

The traffic runs both ways, incidentally. カラオケ (karaoke) was built in Japan from 空 (kara, "empty") and オーケストラ (ookesutora, "orchestra"), and it went the other direction into English intact.

Reading a katakana word you have never seen

The order that wastes least time:

1. Read it as romaji. Slow, syllable by syllable. ストロベリー is su-to-ro-be-rii.
2. Restore what English swallowed. Drop the vowels that Japanese added to carry final consonants and rejoin what the syllables split apart. sutoroberii collapses to "strawberry".
3. If nothing appears, suspect an abbreviation. Six kana or fewer with no resemblance to an English word is usually a stub — パソコン, アニメ.
4. If it looks like English but the meaning is wrong, it is 和製英語. Trust the Japanese meaning, not the English pieces.
5. Otherwise it may not be from English at all. アルバイト (German), パン (pan, "bread", from Portuguese), エネルギー (enerugii, from German Energie). The older and more everyday the word, the likelier this is.

Step one is the mechanical part, and the romaji converter will do it for you on any katakana you paste in — including the long vowels and small tsu that are easy to misread by eye.

Why this is worth ten minutes

Loanwords are the cheapest vocabulary in the language. You already know the meaning; you are only learning a new spelling of it. A learner who can read katakana at speed recognises thousands of words on sight, before studying any of them.

That makes katakana the one script where learning the characters pays off immediately, rather than after months of vocabulary work. The interactive kana chart has every katakana with its romaji and its sound, and Japanese pronunciation covers the vowel and consonant system these words are being forced into.

Start with the katakana column, and the next menu, sign or product label you meet will have more English in it than you expect.