Arabic Letter Forms: Why Each Letter Has Up to Four Shapes (Complete Chart)

If you've ever tried to match Arabic text letter by letter against an alphabet chart, you've hit the problem quickly: the letters in the word don't look like the letters in the chart. The hāʾ in هو looks nothing like the hāʾ in وجه, and the ʿayn at the start of عين is a different shape from the one at its end.

This isn't a quirk of handwriting. It's the core design of the script. Arabic is always cursive, in print as much as by hand, and almost every letter has several shapes depending on where it sits in a word. Once you understand the system, Arabic stops looking like a wall of squiggles and starts looking like a small set of building blocks.

The four positions

Each letter can appear in up to four positions:

• Isolated: standing alone, not joined to anything.
• Initial: at the start of a connected group, joined only to the letter after it (on its left).
• Medial: in the middle, joined on both sides.
• Final: at the end, joined only to the letter before it (on its right).

Take bāʾ (ب). Alone it's a bowl with a dot below. At the start of a word, the bowl is dropped and it becomes a short tooth: بـ. In the middle it's the same tooth with a connection on both sides: ـبـ. At the end it gets its bowl back: ـب. Put three in a row and you get ببب: initial, medial and final forms, joined.

The important thing for designers is that you never type these forms. You type the plain letter, and the software picks the right shape. That process, called shaping, is why Arabic breaks in apps that don't support it (see our guide to fixing disconnected Arabic text).

The complete chart

The chart below uses the tatweel (ـ), a joining stroke, to show each form on its own. Arabic reads right to left, so in each row the initial form is on the far right of the shapes and the final form is to its left.

The six letters that never connect forward

Look at rows 1, 8, 9, 10, 11 and 27: ا د ذ ر ز و. These six letters only ever join to the letter before them. They never join to the letter after them. So they have just two shapes: isolated (which doubles as the initial form) and final (which doubles as the medial form).

Every one of them forces a break in the word. That's why so many Arabic words visibly split into separate pieces even though the script is "always connected":

• مدرسة (school) = مد + ر + سة. The dāl and the rāʾ both refuse to join forward.
• وزارة (ministry) = و + ز + ا + ر + ة. Almost every letter is a non-connector, so the word is written as five separate pieces.
• أنا (I) = أ + نا. The alif starts a new group, and the nūn joins into the final alif.

These connected groups are what calligraphers and type designers think in. The letters inside one group must flow as a single stroke. The gaps between groups are spacing decisions, and getting that spacing even is a big part of what makes calligraphy look finished.

Letter families: one skeleton, different dots

Strip the dots away and many letters share the same underlying shape, the rasm. The dots (iʿjām) are what tell them apart:

This is why the earliest Qur'an manuscripts, which were often written with few or no dots, are hard for modern readers: the skeleton was there, but much of the information that separates letters wasn't. Our Kufic guide explains how dots were added gradually. In the Maghribi script of North Africa, fāʾ and qāf are still dotted differently today: one dot below for fāʾ and one dot above for qāf.

For learners, the family table is a shortcut. The 28 letters come down to fewer than twenty skeletons. Master those shapes, then learn where the dots go.

The letters that change the most

Most letters keep a recognisable core across their forms. A few change dramatically, and they're the ones that trip up beginners:

• ʿayn (ع): an open, mouth-like head at the start (عـ), a closed, filled triangle in the middle (ـعـ), and the full tailed form at the end (ـع). Compare عمل, بعد and سمع.
• hāʾ (ه): perhaps the most variable letter of all. It has a looped initial (هـ), a two-lobed medial (ـهـ) and a small knotted final (ـه). Calligraphy styles vary these even more.
• kāf (ك): the isolated and final forms carry a small mark inside, which disappears at the start and in the middle (كـ), where the letter becomes a tall slanted stroke.
• yāʾ (ي): two dots below and a deep, swinging tail at the end of a word, but just a small tooth in the middle (ـيـ).

Special characters you'll meet constantly

The 28-letter chart isn't quite the whole story. A few extra characters appear in almost every sentence:

• Tāʾ marbūṭa (ة / ـة): a hāʾ shape with two dots, found only at the end of a word. It usually marks the feminine (مدرسة, جميلة).
• Alif maqṣūra (ى / ـى): a dotless yāʾ found only at the end of a word, pronounced as a long ā (على, مستشفى).
• Hamza (ء) and its "seats": the glottal stop can stand alone (ء) or sit on an alif (أ إ), a wāw (ؤ) or a dotless yāʾ (ئ). Hamza on the line never connects to anything.
• Alif madda (آ): an alif with a wavy stroke above, representing a long ā after a glottal stop (آمن, القرآن).

Lām-alif: the compulsory ligature

When lām is followed by alif, Arabic doesn't draw the two letters side by side. It merges them into one shape, لا. Unlike most ligatures in Latin type, this one isn't optional. Writing ل and ا separately looks simply wrong to an Arabic reader.

It is also the best quick test of whether software is shaping Arabic properly. Type ل then ا. If you don't get لا, shaping is switched off.

Beyond four forms: what calligraphers add

The four-form model is how fonts work at minimum. Calligraphy goes much further.

• Contextual alternates and ligatures. In Naskh and Thuluth, letters change shape depending on which letter they're joining, not just their position. Lām followed by mīm (لم), or bāʾ before yāʾ, each has preferred ligature forms. Richer fonts such as Amiri include hundreds of these, which is a big reason Amiri looks hand-written where simpler fonts look mechanical.
• Stacking. Calligraphers often lift a letter so it sits above the one before it, letting a long word fit a short space. Diwani and Thuluth compositions rely on this heavily.
• Kashida (elongation). A connection can be stretched for emphasis or to justify a line, as in بـــــسم. Good calligraphers stretch only certain connections, and by set proportions. They never stretch every one, the way some justification algorithms do.
• Style-specific forms. Kufic squares every curve, Nastaliq slopes each connected group down to the left, and Ruq'ah shortens or drops some strokes for speed. The six major styles guide shows how each treats the same letters.

How to show a single form in your own designs

Sometimes you want a letter on its own in a specific form, for a logo, a monogram or a teaching sheet.

• Use the tatweel (ـ, U+0640). Type it before or after the letter to force the joined form, as in the chart above. It draws a visible joining stroke, which often looks good in display work.
• Use the zero-width joiner (U+200D). It requests the joined form without any visible stroke. It's handy for a monogram where you want just the initial بـ shape with nothing trailing.
• Pick a font with good isolated-form design. Some display fonts draw their isolated forms carelessly because they're rarely seen. Try a few in the Arabify text studio before committing.

Practice: read the shapes, not the letters

A useful exercise is to take a short word, split it into its connected groups, and name the form of each letter:

كتاب (book) → كتا + ب

• ك: initial · ت: medial · ا: final (and a non-connector, so the group ends) · ب: isolated

سلام (peace) → سلا + م

• س: initial · ل + ا: the lām-alif ligature, ending the group · م: isolated

Do this with five words a day and within a week you'll read connected text by its silhouette. The same habit makes you a sharper designer, because you'll notice at once when software, or a tattoo stencil, gets a shape wrong.

Sources & further reading

• Unicode Consortium, ArabicShaping.txt, the official joining type of every Arabic character (dual-joining, right-joining, non-joining).
• Unicode Consortium, Arabic code chart (U+0600–U+06FF).
• Unicode Technical Committee, comments on Maghribi fāʾ and qāf, on the regional dotting conventions mentioned above.