Arabic Diacritics & Punctuation Codes

Arabic writes consonants and long vowels, leaving short vowels to be supplied by the reader. The optional marks that spell them out are called tashkeel or harakat, and they are the first eight entries in the table above. Like Hebrew niqqud, they are omitted from ordinary writing — newspapers, books, websites and correspondence carry no tashkeel at all — and appear mainly in the Qur'an, in poetry, in children's and learner materials, and wherever a word would otherwise be genuinely ambiguous.

The three basic vowel marks are fatha for a, damma for u and kasra for i. Each has a tanween counterpart — fathatan, dammatan and kasratan — which doubles the mark and adds a final n sound, marking grammatical indefiniteness. Sukun indicates that a consonant carries no vowel at all. Shadda marks a doubled consonant and is the one mark that changes meaning frequently enough that writers sometimes keep it even in unvocalised text: darasa means he studied, while darrasa with a shadda means he taught. The dagger alif is a superscript stroke standing in for a long a that is not written as a full letter, most famously in the word Allah and in هذا.

When a letter carries both a shadda and a vowel mark, the two stack, with the shadda above and the vowel above or below it. This is normal Arabic typography and depends on the font handling mark positioning correctly. Poor font support shows up as marks colliding with the letter or floating in the wrong place, which is a rendering problem rather than a problem with your text.

The final three entries are punctuation and they are not interchangeable with their Latin counterparts. The Arabic comma ، sits high and curves the opposite way to a Latin comma. The Arabic question mark ؟ is mirrored. The Arabic semicolon ؛ is likewise reversed. These are separate Unicode characters, and using Latin punctuation in Arabic text is a visible mistake to Arabic readers as well as one that interacts badly with bidirectional text layout.

Bidirectional text is the other thing to know. Arabic runs right to left, and when it is mixed with Latin text or numbers the display order is computed by the Unicode bidirectional algorithm rather than stored order. Punctuation at a boundary can appear to jump to the wrong end of a line. This is usually the algorithm working as designed on ambiguous input rather than corrupted data, and the fix is directional marks or explicit isolates rather than retyping. On Windows and Linux, tashkeel is entered on an Arabic layout using Shift with the letter keys. On macOS, select the Arabic input source; Option plus a key gives the diacritics. Because these are combining marks, type the consonant first and then the mark.