Technical • 7 min read
One of the first things that surprises people learning to read or design with Urdu or Arabic script is that letters don't have a single, fixed shape the way Latin letters do. The same letter can look noticeably different depending on where it sits in a word — and understanding why is the key to understanding how the entire script functions.
Arabic-script writing is cursive by design — letters within a word are meant to connect into a single continuous flow, similar to handwritten English cursive, except this connecting behavior is built into the script itself rather than being an optional stylistic choice. Because a letter needs to connect smoothly to whatever comes before and after it, each letter has up to four distinct forms: isolated (standing alone, connected to nothing), initial (at the start of a word, connecting only to the letter after it), medial (in the middle, connecting to letters on both sides), and final (at the end, connecting only to the letter before it).
Take the letter ع (ain). In isolation, it has a rounded, almost circular shape with a small tail. At the start of a word, that tail extends rightward to connect to the next letter, changing its overall silhouette. In the middle of a word, it compresses into a smaller, more angular shape sandwiched between its neighbors. At the end of a word, it regains some of its isolated roundness but keeps a connecting stroke leading into it from the right. A reader fluent in the script recognizes all four as "the same letter" instantly, but a learner — or a font missing proper glyph support — can easily render this incorrectly.
A small set of Arabic-script letters — including ا (alef), د (dal), ذ (zal), ر (re), ز (ze), ڑ (rre, specific to Urdu), and و (waw) — never connect to the letter that follows them, even in the middle of a word. These are sometimes called "non-connecting" letters. They still connect to whatever comes before them, but the next letter after one of these always starts fresh, as if beginning a new visual segment. This is why some words appear to have a small gap or break partway through, even though it's still written as a single connected word with no actual space.
In digital typography, this is managed through OpenType contextual substitution — a font contains separate glyph designs for each of a letter's possible forms, and the rendering engine automatically selects the correct one based on the letter's neighbors as it processes the text. This is genuinely more complex than rendering Latin script, which is part of why historically, Arabic and especially Urdu typesetting lagged behind in software support; early digital systems built primarily for Latin text simply didn't have a mechanism for this kind of contextual glyph switching built in from the start.
When you see Urdu or Arabic text rendering with disconnected, isolated-looking letters instead of a flowing connected word, it's almost always because the software or font lacks proper OpenType shaping support — the underlying Unicode text is correct, but the rendering engine is displaying each character's isolated form instead of applying the contextual substitution rules.
Nastaliq pushes this contextual complexity even further than Naskh, because letters don't just change shape — they also shift vertically and diagonally relative to each other as a word cascades down and to the left. A well-built Nastaliq font needs not only the four standard contextual forms for each letter, but also rules for how much vertical offset to apply based on the specific combination of surrounding letters. This is a major reason why high-quality digital Nastaliq fonts took so much longer to develop than Naskh fonts, and why even today, the number of genuinely well-engineered Nastaliq fonts available is much smaller than the number of Naskh fonts.
You can observe contextual letter forms directly using our Urdu Alphabet Explorer, which shows each letter alongside an example word, or by typing different words into our Calligraphy Generator and noticing how the same letter looks different depending on where it falls in each word.