Typing Amharic and Tigrinya: a syllable at a time

One character is a consonant and a vowel

Ethiopic — also called Ge'ez script, or fidäl — is an abugida, which means its unit of writing is a syllable rather than a sound. A character does not say "b". It says "bu", and a different character says "bi", and another says "be".

The arithmetic follows immediately. Around 33 base consonants, each appearing in seven vowel forms, gives over two hundred characters before you count the extras that particular languages need — and more than 300 are in common use.

That is not a large alphabet. It is a different kind of thing from an alphabet, and comparing it to one is where most confusion about typing it begins.

Which is why there is no keyboard chart here

Every other script on this site can be drawn as a keyboard, because a key produces a character. Ethiopic cannot, and the reason is structural rather than a gap in the data.

You type a consonant and then a vowel, and the input method replaces the pair with the single character that means both. So a key does not produce a finished character — it produces half of one. A chart of single keypresses would show you the consonants and be silent about every character you actually want, which is worse than showing nothing and saying why.

The scale of the problem is visible in the practice words themselves. The table below counts the distinct characters each language's own practice words need, and Ethiopic sits where you would expect: a few dozen words already ask for more distinct characters than a keyboard has keys, while a hundred English words never leave the 26 letters.

The lessons here drill those characters directly rather than teaching key positions, which is the honest response to a script where the key positions are not the skill.

Closer to an input method than to a layout

Typing Ethiopic is therefore closer to typing Japanese or Chinese than to typing Greek, even though nothing about the result is exotic: Ethiopic runs left to right, uses ordinary word spacing in modern text, and needs no conversion step or candidate list.

What it shares with the CJK methods is the composition. Two keystrokes become one character, the character appears once the second one lands, and what you press is not what you see. What it does not share is ambiguity — there is exactly one character the pair can make, nothing is offered for you to choose, and nothing depends on context.

That makes it much easier than it first sounds. The rule is regular across the whole script, so learning it is learning one pattern rather than three hundred positions.

It also means touch typing transfers almost completely. The keys are the ones under your fingers now, and the consonant half of every syllable is an ordinary Latin-key reach.

Its own punctuation, and its own numbers

Ethiopic brought more than letters. It has its own numerals, and its own punctuation — including a word separator written as two vertically stacked dots, which historically did the job a space does.

Modern text generally uses ordinary spaces, so this is a thing to recognise rather than to type. It is worth knowing because it turns up in older material and in formal settings, and because it explains why Ethiopic text sometimes looks differently spaced from what you expect.

Tigrinya is written in the same script and typed the same way. It uses several characters Amharic does not, reflecting sounds Amharic lost, so an Amharic keyboard is close to sufficient for Tigrinya and not quite — and software support for Tigrinya specifically is thinner, which is a smaller version of the problem that affects most languages with fewer than a few tens of millions of speakers.