When a number, an English name, or a date sits inside an Arabic sentence, it follows its own direction rules within the surrounding right-to-left flow — not a simple mirror of the rest of the line.
Sourced from the W3C's Arabic & Persian Layout Requirements and the Unicode Bidirectional Algorithm — reviewed by Ata Rafi, Founder and Chief Formatter at Format Fix.
Request A Free QuoteArabic text flows right-to-left, but numbers and Latin-script words embedded in it behave as their own left-to-right "runs" positioned within that right-to-left line. Punctuation and other neutral characters take their direction from whatever strong-direction text sits next to them. This is governed by the Unicode Bidirectional Algorithm, not a font setting or a manual flip.
Numbers are written with the lowest-significant digit on the right and the highest on the left — left-to-right — even though the surrounding Arabic script reads right-to-left. A multi-digit number keeps its normal reading order; only its position within the line follows the RTL flow.
The digit sequence itself reads normally left-to-right; only its placement within the RTL sentence shifts based on where it appears.
A tool applies the paragraph's RTL logic to the digit sequence itself instead of treating it as its own embedded left-to-right unit, reversing the digit order. See Fixing Arabic Text Direction in InDesign for a documented case of this exact bug.
Currency symbols are treated as part of the number and placed to its left; degree symbols are placed to its right, with a separating space required between a quantity and its unit.
An embedded English name reads left-to-right internally, exactly as it would in an English sentence, but its position within the surrounding right-to-left line follows the bidi algorithm's placement rules rather than simple mirroring. This is one of the least-documented cases in Arabic typesetting standards — even the W3C's own Arabic & Persian Layout Requirements has no dedicated guidance for personal names in mixed-script text, leaving it to be worked out case by case in practice.
A manual character-level direction override applied to "fix" how the name looks, which can conflict with the automatic bidi placement and create a different, less predictable error than the one it was meant to solve.
Check the name's position relative to the surrounding Arabic words after any edit to the sentence, not just once when it's first typed — a change elsewhere in the line can shift how the name sits.
Let the bidi algorithm position the name automatically rather than manually overriding direction at the character level, and re-check placement after any surrounding edit rather than assuming it holds.
A date's digits follow the same left-to-right internal reading order as any other embedded number, positioned within the RTL sentence around it. The most common failure isn't the date itself — it's a numeral system that doesn't match the rest of the document, switching between European and Arabic-Indic digits inconsistently across a single manuscript.
Punctuation touching an embedded English word or number follows the same placement conventions as Arabic punctuation elsewhere in the sentence — it doesn't get special treatment just because it's adjacent to Latin script. See Arabic Punctuation Placement Problems for the specific rules and documented bugs around comma, period, and question mark placement.
This page covers what happens within a single line when Arabic and English mix. A fully bilingual book is a different scale of problem: binding direction, shared or separate page-numbering sequences, and cross-script footnotes across an entire manuscript.
Line-level: how a number, name, or date behaves when embedded in Arabic running text.
Book-level: binding direction, page-numbering sequences, and cross-script footnotes across an entire manuscript. See Bilingual Arabic-English Book Formatting.
Examples of manuscripts with heavy mixed-script content are in the portfolio, or get a quote for a project with this kind of content already in progress.
Writers whose Arabic manuscript references English-language sources, authors, or terms inline throughout the text.
Scholarly manuscripts with heavy embedded citations, dates, and figures mixed directly into Arabic running text.
Translators moving content between English and Arabic who need embedded numbers, names, and dates to read correctly in the final layout.
Arabic text flows right-to-left, but numbers and Latin-script words embedded in it behave as their own left-to-right runs positioned within that line, governed by the Unicode Bidirectional Algorithm rather than a font setting or manual flip.
Numbers keep their normal left-to-right digit order, with the lowest-significant digit on the right, even though the surrounding Arabic script reads right-to-left. Only the number's position within the line follows the RTL flow.
It depends on the region: European numerals are common in Western Arabic-speaking regions, Arabic-Indic numerals in Eastern Arab countries, and Eastern Arabic-Indic numerals in Iran and Afghanistan. The most common error is mixing systems inconsistently within one document.
The name reads left-to-right internally as normal, but its position in the surrounding RTL line follows bidi placement rules rather than simple mirroring. This case is genuinely underdocumented in Arabic typesetting standards, and manual character-level overrides intended to "fix" it often introduce a different error instead.
A date's digits follow the same left-to-right internal order as any embedded number. The most common mistake is a numeral system that doesn't match the rest of the document, not the date format itself.
No — this covers line-level behavior of embedded numbers, names, and dates. A fully bilingual book involves book-level decisions like binding direction, page-numbering sequences, and cross-script footnotes across the entire manuscript.
Verify numeral order and system consistency throughout, check embedded names in context after edits, match date formatting to the rest of the document, and confirm punctuation around embedded English words and numbers follows the correct Arabic convention.
Related: how Arabic dialects and regional differences affect formatting.
