Greek How-To

Please note that submission in plain ASCII is no longer required by Project Gutenberg. This information may still be useful, however. This How-To was written in days when Project Gutenberg required submissions in plain ASCII.

Transcription of English texts into ASCII is relatively straightforward; however, you may sometimes encounter sections of text which ASCII was never designed to represent. Greek text is one of the more common examples, especially in public-domain texts. Greek and Latin were much more common among the literati even as late as the beginning of the 20th century. Fortunately, the use of Greek in English texts is usually limited to a few words, or a memorable quote.

You have several options in the treatment of such text. You can, of course, simply replace the Greek with [Greek text] or a similar declaration. But with a little more work, you can faithfully represent the Greek text in a manner both effective and globally accepted.

The Greek Alphabet

Western languages that use the Latin alphabet (such as English) have long accepted the transliteration of Greek into Latin characters in the following manner:

Numerical value Letter Uppercase Lowercase Transliteration
1 alpha Α α a
2 beta Β β b
3 gamma Γ γ g1
4 delta Δ δ d
5 epsilon Ε ε e
7 zeta Ζ ζ z
8 eta Η η ê or e
9 theta Θ θ th
10 iota Ι ι i
20 kappa Κ κ k
30 lambda Λ λ l
40 mu Μ μ m
50 nu Ν ν n
60 xi Ξ ξ x
70 omicron Ο ο o
80 pi Π π p
100 rho Ρ ρ r
200 sigma Σ σ s
  final sigma (at end of word) ς    
300 tau Τ τ t
400 upsilon Υ υ u or y
500 phi Φ φ ph
600 chi Χ χ ch
700 psi Ψ ψ ps
800 omega Ω ω ô or o

(1) Gamma. When gamma is found in combination with another consonant, it may be transliterated as an “n”: γγ = ng; γκ = nk; γξ = nx; γχ = nch.

For the most part, transliteration is straightforward. There are no combinations that transliterate differently in different contexts, unless you are translating from Greek to some other language. Most modern languages, for example, use only one form of a person’s name, where the Greek may have several different forms. As you can see above, “upsilon” can be rendered as an English “u” or “y”. Generally, if upsilon follows another vowel, like alpha, epsilon, or omicron, use “u”. Otherwise, use “y”. This rule isn’t set in stone, but it’s clear and simple.

Update: It has recently come to my attention that some texts use the even more obscure ligature “ou”, which looks like this (lowercase): Greek ou ligature. Feel free to transliterate this as “ou” wherever you see it.

Accent-marks (diacriticals)

The only tricky bit is the rough-breathing mark which is usually transliterated as an “h” preceding the letter in question. `a = ha The exception is a word that starts with “r”, which is begun with “rh” instead (e.g. the letter “rho”). It can get sneaky when combined with other diacriticals (accent-marks), especially the “soft-breathing” mark, which does not get transliterated at all.

Rough-breathing Soft-breathing
Greek dasia.gif Greek psili.gif

The rough-breathing mark may appear above or in front of the initial letter. It’s only found at the beginning of a word. Some example text:

Mênin aeide, thea, Pêlêiadeô Achilêos

Mênin aeide, thea, Pêlêiadeô Achilêos

oulomenên, hê muri' Achaiois alge' ethêke

oulomenên, hê muri’ Achaiois alge’ ethêke,

Notice that “Achilles” and (the) “Achaeans” (this is the opening of the Iliad) have the silent-breath mark, so they don’t get the “h”; but the word “he” on line 2 does, since it has the rough-breathing mark. Also, notice the eta and omega characters: we’ve gone ahead and used the e and o characters with circumflexes. These are Unicode characters 234 (ê) and 244 (ô). The Distributed Proofreaders formatting guidelines say to use the accented characters when possible. When accented characters are not available, use “e” for both eta and epsilon, and “o” for both omicron and omega. Don’t let the other diacriticals (accent marks) throw you: nobody knows for sure what they mean, and for the most part, they don’t affect the transliteration. Just worry about the rough-breathing mark. Finally, notice that I gave “myri” and “alge” the single-quote character at the end; this mark shows that Homer left off some letters to improve the meter. They function like the apostrophe in English contractions.

Translators and printers of Ancient Greek are always guessing about punctuation.


The original transcribers use their knowledge of Greek phrasing to determine where to place punctuation marks. There are usually two punctuation marks that differ from English: a question mark, and a medium stop. Modern printings of Greek often use a semicolon “;” to indicate a question. Use a question mark “?” instead. A single dot usually represents a medium stop; replace that with an English semicolon “;”.


Ancient Greek used letters to indicate numbers. So you can transliterate ΙΛΙΑΔΟΣ Α as ILIADOS A or ILIADOS 1. And yes, there are letters which fell out of favor, but still had numeric value: digamma (or stigma) = 6, koppa = 90, sampi = 900. You will probably never see these in practice.

Thanks to Stavros Macrakis for his suggestions in updating this document.

Happy transliteration! Written May 1999 by Robert Brewer.

Updated: June 2004.