Transliteration Rules

What is Transliteration?

Deep Informational Technologies (Deep IT) have built Transliteration for Mac OS (only). As you may have imagined already, this software does exactly what it says on the tin: it transliterates! That is, it maps strings of text from one writing system to another (e.g. from Greek alphabet to Latin).

Download it here!

Kinds of transliteration:

There are various kinds of transliteration, e.g. phonetic transliteration preserves the sound of the source text, while orthographic transliteration preserves the spelling of the source. That means that, depending on the kind of transliteration, if the transliterated text is “backwards” transliterated it may result to a text slightly different from the original.

Direction:

For example, if we transliterate both “η” (Greek eta) and “ι” (Greek iota) to “i” in order to keep the phonetic value of the letters, it is difficult to decide how we’ll transliterate “i” back to Greek. There are no particular rules to decide when eta and when iota is used in Modern Greek, because the spelling of each word depends on its history (words that date back to Classical Greek vs modern words).

For this reason, it is important to keep in mind to use the original direction of the transliteration. If the rule is named, e.g. Greek→Latin, expect the transliteration to work from Greek to Latin, and not vice versa. As a matter of fact, Latin to Greek results may be legible to advanced readers of Greek; however, never use such results for work. You might get away with it in informal documents though (emails, etc).

In any case, it is obviously best practice to consult a fluent reader of both source and destination writing systems before using transliterated text.

My rules:

Having used and liked this software, it was more than natural to contribute some rules for it. The rules are built in an XML file in the PLIST format. Below you can find information about each set:

Greek→Latin

Greek to Latin (results may have been improved since the screenshot was taken)

Greek is one of the languages that is particularly difficult to transliterate. You can see various romanization rules of Greek. However, some of them only apply to Ancient, Classical Greek.

I have built rules to transliterate Modern Greek into Latin, using “orthographic Greeklish” rules.
These rules apply to all Greek dialects, e.g. Cypriot, Pontic, etc.

The rules have been based on both ISO 843 and UN V/19 systems, which are variations of the ELOT 743 systems (see relevant United Nations report).

Ladino→Hebrew

Ladino to Hebrew (results may have been improved since the screenshot was taken)

The Hebrew-Latin rules that already exist in Transliteration software serve traditional Hebrew very well; however, other languages use Hebrew script, too (Yiddish, Ladino, Yevanic, etc). Not all of them are served equally well.

For this reason I built transliteration rules for Ladino (also called Judaeo-Spanish). I used the “phonetic transcription” rules proposed by the Israeli Autoridad Nasionala del Ladino (the official body of the Ladino language) and these notes from Lehrhaus Judaica, Berkeley, CA.

Please note that Ladino in Hebrew characters is read from right to left.

Armenian

coming soon…

Georgian→Latin

I completed rules to transliterate the Georgian alphabet to Latin.

This time, I used a combination of orthographic and phonetic transliteration.  This was achieved through heuristic evaluation with native speakers and through monitoring social media. One could say the result is a simplified mix of the Georgian government’s proposed romanisation and ISO9984.

Georgian to Latin (results may have been improved since the screenshot was taken)