Typing the Arabic script: one alphabet, five keyboards

The script is shared, the keyboards are not

It is easy to assume that languages written in the same alphabet type the same way. For the Arabic script it holds for two of the four languages that join Arabic here and fails completely for the other two, and the difference is not a matter of degree.

Persian and Pashto are built on the Arabic arrangement. They move fewer than a third of the keys, mostly to make room for letters Arabic does not have, and an Arabic typist sitting down at either is immediately most of the way there. Urdu and Kurdish are not variants of the Arabic layout at all — they are separate arrangements that happen to write the same letters, and knowing Arabic buys you almost nothing on them.

The count below is over the keys both layouts define, which is the three rows of letters a keyboard chart draws. It is measured rather than characterised, so it moves if the layout data ever does.

Why the letter you just typed changes shape

Arabic letters take four forms depending on where they fall in a word — isolated, initial, medial, final — and most of them join to the letter that follows. This is the single most disorienting thing about learning to type the script, and it is worth understanding before it happens rather than during.

You press one key per letter. The font and the text renderer choose the form, and they revise their choice as the word grows, so a letter you typed a moment ago visibly changes when you type the next one. Nothing has gone wrong and you have not pressed the wrong key. It is also why Arabic-script typists learn to watch the screen rather than their hands earlier than most: the screen is telling you something the keyboard cannot.

It has a practical consequence for any keyboard chart, including the one on these lessons. A chart has to show one form per key, and the form it shows is the isolated one — the letter standing alone. Inside a word that same letter will usually look different. The key is right; the shape is context.

What each language adds

Every language here writes sounds Arabic does not, and each solved that the same way: by adding letters. This is where the extra keys in the table above go.

The additions are not interchangeable, which is the reason five keyboards exist rather than one with everything on it. A layout carrying every letter of all five languages would need far more keys than a keyboard has.

Persian adds four letters

پ, چ, ژ and گ, for sounds Arabic lacks. It also writes two shared letters differently: Persian ک and ی have their own standard forms, and using the Arabic ك and ي instead is a common and visible error in Persian text.

Urdu adds retroflex consonants and an aspiration mark

ٹ, ڈ and ڑ for the retroflex sounds, and do-chashmi he — ھ — which marks aspiration rather than standing as a letter of its own. It also has ے and ں, neither of which Arabic uses.

Pashto adds the most

ټ, ډ, ړ, ږ and ڼ among others, most of them written by putting a small circle beneath an existing letter. Around 45 letters in total, which is why the standard Afghan layout leans so heavily on shifted positions.

Kurdish writes its vowels

Sorani marks vowels explicitly, which Arabic does not — a deliberate departure that makes the script easier to read for a learner. That needs letters of its own: ە, ۆ and ێ, alongside ڤ, ڵ and ڕ.

The invisible character Persian cannot do without

Persian uses the zero-width non-joiner constantly. It is often called the half-space, and it separates the parts of a compound word so they do not join visually while staying one word.

The thing to understand about it is that it is required for correct spelling in a great many Persian words, and it is invisible. Text typed without it is not merely untidy: it looks wrong to a Persian reader, and it is difficult to search, because a word written with a real space is a different string from the same word written with a half-space.

The standard Persian keyboard gives it a key. It does not appear on the three letter rows a keyboard chart draws, which is part of why it goes unmentioned so often — and it is the single most important thing about typing Persian that nobody tells you.

Which keyboard knowledge transfers

If you already type Arabic, Persian and Pashto are a short walk. Persian changes nine of the twenty-eight keys it shares with the Arabic layout and Pashto changes eight, and in both cases the changes cluster around the letters those languages added. The rest of your recall is intact.

Urdu is a different keyboard. It differs from the Arabic layout on twenty-four of the twenty-seven keys both define, which in practice means you are learning it from scratch rather than adjusting. If you have read anywhere that Urdu uses a phonetic layout — where each letter sits on the Latin key that sounds like it — that describes one of the alternative Urdu keyboards in circulation, not the standard one these lessons teach.

Kurdish Sorani is the interesting case, because it is the one arrangement in this family that really is aligned to Latin. Twenty-four of its twenty-six letter keys carry the letter a reader of Kurdish's own Latin alphabet would predict, including the three that would trip up an English speaker: x is خ, c is ج and j is ژ, which are exactly their values in Kurdish Latin. Someone who reads Kurmanji can nearly guess this keyboard.

Across all five, the physical half of touch typing transfers completely and immediately. What does not is the map — and the map is what the practice is for.