Spelling Foreign Words and Names in Hebrew

Hebrew absorbs foreign words at speed — technology, food, business, pop culture — and every one of them has to be respelled in twenty-two letters that were never designed for the job. There is a governing principle, and internalising it saves you from the single most common learner error:

Hebrew transcribes sound, not spelling. The source language's own orthography is irrelevant. Thomas is not respelled letter by letter; it is listened to and written down as תומס. Knight would be נייט. What English wrote is invisible; what English says is what gets recorded.

The consonants Hebrew did not have

Three sounds arrived with modern loanwords and got a geresh ׳ to carry them, borrowed from Yiddish and Judeo-Arabic practice:

The rest of the consonants are matched by ear, and four choices are worth memorising because English speakers get them wrong in a predictable direction.

The last row is not a choice at all: Hebrew has one letter for both p and f, so פ does double duty and only your knowledge of the word tells you which it is. See בגדכפ״ת for why.

Then come the two systematic exceptions, and they are systematic because they come from Greek. Where the source has Greek theta or Latin th, Hebrew writes ת; where it has Greek chi, Hebrew writes כ:

This is why technology is טכנולוגיה with a כ while classic is קלאסי with a ק, although both are pronounced with a plain k. The spelling preserves the word's route into Hebrew, not its sound — the one place where the sound-first principle bends.

v, w and the doubled ו

Hebrew has no w sound, so every English w becomes v: Washington is vashíngton, William is vílyam, whisky is víski. There is nothing to reach for; the sound simply does not exist in the language.

Writing it is the fiddly part. Because ו is also the vowel o/u, a consonantal ו in the middle of a word is doubled — טלוויזיה, סוודר — exactly as in native words like תקווה. At the start of a word the Academy's rule keeps it single (ויסקי, ונציה), but real usage very often doubles it there too, which is why you will meet both וושינגטון and ושינגטון in print. That dispute is live and it is covered on contested spellings.

The vowels

Unpointed Hebrew can only write three of the five vowels, and loanwords follow the same ktiv male rules as native ones — just more generously.

The row that trips people is the third one. a and e are invisible in Hebrew, native or foreign — טלפון has a vowel between ט and ל that nothing on the page records. Learners often want to insert a י for the e because they have seen it in אינטרנט; but that י is marking the i of in-, not an e. Two consonants in a row with no ו and no י between them are usually separated by an a or an e, and you supply it from your knowledge of the word.

Names, and the split that surprises everyone

Personal names normally get transcribed by ear: Jack is ג׳ק, Michelle is מישל, Steve is סטיב.

But a name with a biblical original is not transcribed at all — it is translated back. Hebrew already owns the name, so the ancestor's form is used for the ancient figure and a modern transcription for the modern person. The result is that one English name maps onto two Hebrew ones, and which you pick declares who you mean:

The same logic runs through place names. Countries and cities that Hebrew has known for two thousand years keep their Hebrew names rather than a transcription: Spain is ספרד, France is צרפת, Greece is יוון, Egypt is מצרים. Newer ones are transcribed: קנדה, ברזיל, תאילנד. See country names.

Who decides

The Academy of the Hebrew Language publishes formal transcription rules — כללי התעתיק — and they are largely ignored outside academic and cartographic work. The working standard is what the press and Hebrew Wikipedia do, and those do not always agree with each other.

The honest consequence: for many foreign names several spellings circulate at once and none of them is simply wrong. Search a news site for a footballer's name and you will find two versions in the same week. Read generously, and when you write, copy the spelling used by the outlet or the person concerned.

Common Mistakes

1. Transcribing the English spelling instead of the sound. The classic case is silent or digraph letters.

2. Using ת for an ordinary t. ת is reserved for th in the source; a plain t takes ט.

3. Using כ for a k sound in an everyday loanword. כ is either kh or a Greek-derived k; ordinary k is ק.

4. Forgetting the geresh on a final ג or צ. Without it the letter is a plain g or ts.

5. Carrying a doubled consonant across. Hebrew simplifies it.

Key Takeaways

• Hebrew transcribes sound, not source spelling — Thomas → תומס, Carl → קרל.
• ג׳ = j, צ׳ = tsh, ז׳ = zh; these three sounds exist only in loanwords and only with a geresh.
• A plain t is ט, a plain k is ק; ת and כ are reserved for words that came through Greek (מתמטיקה, טכנולוגיה).
• o/u take ו, i takes י, a and e are written with nothing, an initial vowel takes א and a final a takes ה.
• Hebrew has no w: every English w becomes v, written ו — doubled וו inside a word, disputed at the start.
• Consonant doubling is dropped: אנה, וילה, פרארי.
• A name with a biblical original is translated, not transcribed — מיכאל for the archangel, מייקל for the American.
• No single authority wins. The Academy's transcription rules exist; press and Wikipedia usage is what people actually follow, and several spellings of one foreign name coexist.