Right-to-Left Reading and Mixed-Direction Text

Persian words and sentences run from right to left, but not every item inside a Persian line reverses. Numbers keep their own internal order, and a Latin expression such as an email address remains left to right. Learning to separate reading direction, character storage order, and screen layout prevents most early confusion.

These principles are register-neutral: the same direction rules apply in a novel (literary), a government form (formal), a university article (academic), and a text message (informal). What changes across registers is vocabulary and style, not the underlying direction of the script.

Start at the right edge

The first Persian word of a normal line appears at the right. Continue leftward word by word. Within each word, the first letter is likewise the rightmost letter. Thus کتاب ketâb is logically the sequence ک + ت + ا + ب, even though the completed word occupies space from right toward left.

In the second example, begin with من at the right, then فارسی, then می‌خوانم, and finally the period at the left edge. English readers often move their eyes correctly over a Persian word but then jump back to the wrong side for the next word. Keep one consistent right-to-left sweep across the whole clause.

Logical order is not visual order

Computers store text in logical order: the order in which the characters are read. You type ک first, then ت, then ا, then ب. A Unicode-aware program applies the bidirectional algorithm and displays the word right to left. You should never type the last visible letter first to compensate for the screen.

This distinction matters when moving a cursor. Depending on the editor, the visual left and right arrow keys may follow screen position, while deletion acts on the next logical character. Selection can also look discontinuous when a line contains both Persian and Latin. That is layout behavior, not a reversal of the stored Persian word.

The same principle explains letter shapes. A font chooses an isolated, initial, medial, or final glyph from the neighboring characters; the typist enters the basic letter once. The joining forms page shows how this works.

Punctuation follows the Persian line

Sentence-final punctuation belongs at the visual left end of a Persian sentence. Persian uses ؟ for a question mark and ، for a comma. A full stop has the familiar shape ., but its location follows the right-to-left sentence. The Persian semicolon is ؛.

The romanization is an English-direction reading aid, so it uses an ordinary comma and question mark. That does not license those Latin punctuation marks inside Persian target text. A dedicated Persian punctuation page covers typography in more depth.

Numbers are left-to-right islands

Persian sentences flow right to left, while the digits inside a multi-digit number keep their familiar place-value order. In ۱۴۰۵, the thousands digit ۱ remains the first digit of the number; you read the numeral as one left-to-right unit embedded in the larger right-to-left line. Do not reverse the digits when typing.

Punctuation around a number can make cursor behavior surprising, especially with parentheses or a colon. Read the whole number as an island, then resume the Persian line to its left. Persian target text uses Persian digits ۰۱۲۳۴۵۶۷۸۹ (written-standard); the separate Persian digits page teaches their shapes and naming.

Latin text inside a Persian line

Latin words, model numbers, URLs, and email addresses also form left-to-right islands. In an HTML page or modern editor, Unicode normally lays them out automatically. For example, the logical string represented below begins with the Persian word on the right, contains an LTR product name, and then resumes Persian:

Mixed text can look different when copied into a weak editor because neutral characters—spaces, parentheses, slashes, hyphens, and punctuation—take their direction from context. The underlying character order may still be correct. Avoid “fixing” a display problem by reversing letters or adding random spaces. In publishing or software, direction-aware markup provides the reliable solution.

Spaces, ZWNJs, and visible gaps

A visible gap does not always mean “new word.” An ordinary space is a true word boundary. A ZWNJ, or half-space, is an invisible code point that prevents joining inside one orthographic word, as in می‌روم and خانه‌ها. Finally, one of the seven non-connecting letters can create a visible gap inside a word without any space character at all.

The gap inside دانشگاه after a non-connecting letter is not a boundary; the ZWNJ in می‌رود is a controlled internal boundary; the spaces around به separate words. Their appearance may be similar, but their data and grammatical roles differ. See The Seven Non-Connecting Letters before using appearance alone to divide words.

Common Mistakes

1. Entering a word backward to make it “look RTL”

2. Entering sentence-final punctuation before the sentence

3. Reversing the digits of an embedded number

4. Using a Latin question mark in Persian text

5. Treating every visible gap as a space

Key Takeaways

• Read Persian lines from the right edge toward the left, and type letters in logical reading order.
• Fonts handle visual reordering and contextual joining; do not reverse the stored letters yourself.
• Persian punctuation follows the RTL sentence, while numbers and Latin strings are internally LTR islands.
• A visual gap may come from a space, a ZWNJ, or a non-connecting letter; those are three different structures.