← — Leftwards Arrow

Code point 8592 · U+2190 · directional

← is the named character reference for the leftwards arrow, code point 8592. Written as a decimal reference it is ←; as a hexadecimal reference, ←. All three resolve to the same glyph, and the choice between them is about what survives your toolchain rather than about meaning.

Which form to write

Named references are readable and self-documenting, which is why hand-written markup uses them. Numeric references — decimal or hexadecimal — are unambiguous and immune to entity-table gaps, which is why generators, serialisers and sanitisers prefer them. HTML5 recognises the named form in both HTML and XHTML syntax, and browsers resolve any of the three identically at render time.

What this site's encoder does with it

Paste the leftwards arrow into the encoder and pick a style: with named selected the output is ←. Note the fallback: the encoder only has named forms for the four markup-critical characters, so a character that does have a named entity still comes out as a hexadecimal reference here. That is deliberate — the output stays valid even when a downstream parser does not know the name.

This character never has to be escaped in HTML text. Escaping it is a style decision — useful when the source must stay pure ASCII, or when the file travels through systems that assume a Latin-1 encoding.

Mistakes worth avoiding

• Double escaping. Writing ← in source produces the literal text ← on the page instead of the character. Escape once, at the boundary where data becomes markup.
• Missing semicolon. Browsers accept some named references without the trailing semicolon in text, then swallow the following character into the reference name when it is alphanumeric.
• Escaping inside <script>. Entity references are not decoded in script or style content, so an escaped value there stays literal and breaks the code rather than the markup.

Questions about &larr;

What is &larr; in decimal and hex?

Decimal ←, hexadecimal ←. All three reference code point 8592 and render the same character, “←”.

Do I have to use the named form?

Nothing compels the named form. Numeric references work everywhere and survive editors that mangle entity names, which is why minifiers and escape tools emit them by default.

Which form does this site's encoder output for &larr;?

It emits ← into the escaped output. Only &, <, > and " have named forms in the encoder, so every other character is written as a hexadecimal numeric reference even when a named entity exists.

Where does &larr; break?

It breaks when the surrounding text already assumes an ASCII-only pipeline. Some legacy systems re-encode the ampersand, turning one entity into two and exposing the raw code in the rendered page.