Greek transliteration is all fun and games

by Chris Alden

Paphos or Pafos? Aigina or Aegina? Writing Greek placenames into English can cause a world of confusion, says Chris Alden

If you're like most people, your first brush with the perils of Greek-English transliteration will come when you're driving along a Greek or Cypriot road, late at night, looking for a hotel. You know, because your guidebook says so, that the hotel is in the city of Hania. So why on earth do the signs say Chania? Is it the same place? Jet-lagged and impatient, you'll probably take a punt and decide that it is - but until you wake up next morning and look out of the hotel window, you won't be 100% convinced that you've got it right.

Think of a modern Greek place name, and chances are you can work out more than one way of spelling it in English. You can buy an airline ticket to the Cypriot resort of Paphos, but when you get there, the signs will say "Welcome to Pafos". English speakers curse the traffic in Herakleion, but Cretan road signs may point to "Iraklio" (among other possible spellings). You can follow signs to the village of Anogyra and, once there, eat at the Anoyira Tavern. Same place, same name in Greek - more than one spelling in English. What's a Greek learner to do?

The reason for the confusion, of course, is that when it comes to transliterating Greek into English, there's more than one way of skinning the cat.

The only solution, short of ignoring the problem and hoping it goes away, is to try to understand them all.

To simplify things a bit (OK, a lot), it's possible to identify three main lines of thinking in Greek-English transliteration. Broadly speaking there's classical transliteration, which is based on historical academic descriptions of ancient Greek; phonetic transliteration, which attempts to approximate the sounds of Greek today; and a modern-day half-way house, official transliteration, loved by Cypriot signmakers and widely abused by everyone else.

Classical transliteration is the system that's historically been used to render ancient Greek words into English - be they words used in science or medicine, place names, or the names of the letters themselves. One feature of the system is that consonants are "harder" than their modern Greek equivalents, so β=b, γ=g, δ=d, π=p and so on. You'll recognise a large number of double-letter forms from Greek-derived English words: for example, φ=ph (as in phoneme), χ=ch (as in character), etc. As for the vowels, the most repeated differences from modern transliteration systems are first, that η=e; and second, that rough breathings over initial vowels are transliterated with an initial letter h (as in Herakleion). In British English, interestingly, a couple of vowel digrams from ancient Greek are preserved: αι=ae (as in haematoma) and οι=oe (as in oedema), while American English has them both as just "e".

Why does the classical system matter? Because certain Greek-derived spellings have been used in English for centuries - and have now pretty much stuck. For example, English speakers have been spelling Πάφος as Paphos since the year dot, so they aren't about to change now, whatever the local ministry of works may say. Similarly, English speakers will carry on calling the second letter of the Greek alphabet beta to the end of their days, even if in modern Greek it's pronounced like a v.

The problem, of course, is that it isn't describing the modern language - it's describing (or attempting to describe) the ancient.

The "modern phonetic" scheme is about as far from classical transliteration as you can get. This describes the attempt, or rather attempts, to transliterate Greek by echoing modern Greek pronunciation. Theoretically, this sounds like a great idea. It means that vowels aren't mangled too much, and redundant symbols such as breathings (which aren't even spelt, let alone pronounced, in modern Greek) are completely ignored - so that ultimately, if you read out a place name spelt under this system, a Greek will have a fair chance of understanding what you're trying to say.

But of course, Greek isn't English - and Greek letters don't map to English letters all that precisely. As is fairly well known, η, ι, οι, υ and ει are all pronounced i in modern Greek; but in some phrasebooks you'll also see ee - as the real sound is half way between the two. Similarly, π is somewhere between a b and a p, so which one do you use? How do you tell apart the "soft th" that is δ and the "hard th" that is θ (some go with dh for δ, but is that really phonetic at all?) Is it simplest to render χ as h or use kh or ch (its real pronunciation is often described as "like the ch in loch"). And as for the complete nightmare that is γ - pronounced like the consonant y before some vowels, and basically unrenderable in English before others - where do you even start? Is it g or gh or ... what?

Faced with such an explosion of alternative spellings, you can understand where the idea of standardising modern Greek transliteration came about. And if you look at the bilingual road signs in Cyprus (probably the place with the biggest concentration of Greek and English speakers, outside Australia), it's clear that a fairly valiant attempt has been made to create a "official" system that works. These days, most of the old signs in Cyprus have been pulled down and, as far as officialdom is concerned, there is only ever one way to spell "Pafos".

A strange feature of the "official" modern system is that the consonants tend to follow the ancient transliteration scheme quite closely: γ, for example, is rendered as g, and δ=d and π=p, too. Somewhere along the line, though, it seems to have been decided that β=b would be taking things a bit far, so you'll see β rendered as v. The decision to render φ=f, meanwhile (making "Pafos" rather than "Paphos") probably saves a lot of signwriters' time, even if it confuses everyone else. As for the vowels, however, they largely follow modern pronunciation (η=i), and some digraphs are taken letter-by-letter (αι=ai). So it would be Mytilini, not Mytilene, and Aigina, not Aegina, under these rules.

Which is all very well, but the official system is also a huge source of confusion to English speakers. When a certain Greek footballer with "Giannakopoulos" on the back of his shirt plied his trade in Britain, commentators read out his name as if it began with a soft English G - the result, "Jannakopoulos" , sounded awful to the Greek ear; "Yiannakopoulos" would have probably have been more helpful. (It may be the reason he became known as simply "Stelios" - though the length of his name was probably another).

More to the point, people tend to be wedded to their way of transliterating a name, and don't tend to take kindly to changing the way they do it to a new set of rules. If I tell people my name is Yoryos and I eat in the Anoyira Tavern, am I really going to start identifying myself as Georgios from Anogyra?

And if it's difficult to impose standards on those of us who are set in our ways, how much more difficult is to impose them on bilingual, Greek-speaking teenagers, who are in the process of making up the rules all over again, writing as they do in "Greeklish" - a mixture of Greek and English - in text messages and on social media sites? In these systems, the appearance of the symbols is also a consideration: so you'll see η rendered as "h" and θ as "8", alongside all the other contractions and abbrevations typical of textspeak.

As ever with language, it seems the best way forward is just to describe what's going on, rather than try to prescribe a system from the top down - and if you're a writer or an editor, just find a house style and stick to it. The rest of us can revel in the confusion: it's all fun and games in the end, at least once you've found your hotel.

About the writer

Chris Alden is the author of The Greek Alphabet: 24 Letters in 24 Hours and 250 Things to Do in Cyprus on a Sunny Day


Writing systems | Language and languages | Language learning | Pronunciation | Learning vocabulary | Language acquisition | Motivation and reasons to learn languages | Arabic | Basque | Celtic languages | Chinese | English | Esperanto | French | German | Greek | Hebrew | Indonesian | Italian | Japanese | Korean | Latin | Portuguese | Russian | Sign Languages | Spanish | Swedish | Other languages | Minority and endangered languages | Constructed languages (conlangs) | Reviews of language courses and books | Language learning apps | Teaching languages | Languages and careers | Being and becoming bilingual | Language and culture | Language development and disorders | Translation and interpreting | Multilingual websites, databases and coding | History | Travel | Food | Other topics | Spoof articles | How to submit an article


Green Web Hosting - Kualo

Why not share this page:


The Fastest Way to Learn Korean with KoreanClass101

If you like this site and find it useful, you can support it by making a donation via PayPal or Patreon, or by contributing in other ways. Omniglot is how I make my living.


Note: all links on this site to, and are affiliate links. This means I earn a commission if you click on any of them and buy something. So by clicking on these links you can help to support this site.

Get a 30-day Free Trial of Amazon Prime (UK)