An Arabic logotype drawn from the Latin’s own strokes.

A

Solitaire, an upscale women’s boutique, needed one wordmark that worked in English and Arabic together. No Arabic font matched the Latin, so I built the Arabic out of four Latin letters.

The project

Solitaire serves an English- and Arabic-speaking clientele, and the identity had to carry both languages in a single lockup: storefront, labels, packaging, bags. Not two logos for two markets. One logo that held both.

The Latin wordmark is set in Liana, a connected script face by Natalia Vasilyeva for ParaType, with fine hairline joins, generous loops and a light, deliberately feminine hand.

The problem

Liana has no Arabic.

The ordinary answer is to pick the closest available Arabic font and set the name in it. It almost never works. Arabic and Latin type are built on different logic: different writing tools in their history, different placement of contrast, different terminals, a different rhythm along the baseline. Set a stock Arabic face beside a delicate Latin script and the two halves read as two brands that happen to be standing next to each other.

Both scripts here are cursive, which sounds like it ought to help. It does not. A Latin script face joins its letters because the designer chose to; Arabic joins as a rule, each letter changing form according to where it sits in the word. The joins do entirely different jobs.

What I did

Instead of hunting for an Arabic font, I cut up Liana’s actual letterforms and rebuilt Arabic letters from the pieces. The name in Arabic is seven letters. All seven were built from four Latin ones: i, l, g and u.

  • س (sīn), the lowercase u. Sīn is a run of teeth sitting on a continuous baseline stroke, and a script u already carries that: arches joined along the bottom rather than separate marks standing next to each other
  • ي (yā’) and ت (tā’), the same i again, with its dot rebuilt into the pair each letter needs: two below for yā’, two above for tā’
  • ل (lām), the l, with its loop opened out into the straight ascender and baseline curve lām requires
  • و (wāw), the g. Its bowl becomes wāw’s head, its descender becomes the tail
  • ر (rā’), the g again, with the bowl and the closed lower loop cut away, leaving only the descending stroke

Three of those seven letters come from a single Latin one. Arabic builds a whole family of letters by repeating one short tooth and moving the dots, and Liana’s i is a short stroke with a dot above it. Structurally they are the same unit, and finding that is what made the rest possible.

Sīn was the one that took two attempts. The obvious answer is three is with their dots removed, and it works, but three identical repeats look mechanical at the size a logotype gets read. The u gives the same run of teeth with variation built into it, and it joins along the baseline the way sīn actually does.

The dots matter more than they look. In Arabic they are the entire difference between one letter and another, and a dot in the wrong place spells a different word. Liana already carried a dot, drawn in the font’s own hand, so they did not have to be invented. They were moved.

Four things could not move. The Arabic had to be correct Arabic, every letter in its proper positional form, joined where Arabic joins and broken where it does not. The dots had to stay unambiguous, since they are what separate one letter from another. Legibility had to come before style: a letterform a reader has to decode is a failed letterform, however elegant it looks to someone who cannot read it. And the two halves had to sit as equals, sharing a baseline relationship and an optical weight so that neither script dominates and neither looks appended.

What is not in this logo matters as much as what is. There is a well-worn shortcut in which Latin letters are dressed up with hooks and dots to look vaguely Arabic, producing shapes that read as “Arabic” to anyone who cannot read the script and as nonsense to anyone who can. This is the opposite. Latin parts went in; what came out is real Arabic, correctly formed and correctly joined, saying the brand’s name. The vocabulary is borrowed. The grammar is not.

Why it matters

Customers do not read a brand analytically, but they notice this. A carefully drawn Latin wordmark sitting beside a default Arabic font tells an Arabic-reading customer exactly where they came in the priority order, and it does so on the storefront, before anyone has spoken to them.

The shortcut is common because the work sits in a gap. Type designers who could draw the Arabic often cannot read it, and people who read Arabic are rarely in the room when the identity is designed. I read and write Arabic and have been setting right-to-left type since 2005. When a wordmark has to cross scripts, that combination is the whole job.


More on this: Design that survives translation · This smart lock manual shipped with broken Arabic · Portfolio

About the author

Jaber Maycid

Add comment

By Jaber Maycid

Jaber Maycid

About me

I'm a foreign language typesetter and multilingual desktop publisher. Since 2005 I've set type in over 130 languages, including right-to-left and Canadian Indigenous scripts — and I now build automation for the formatting work around translation.

Get in touch →