What _ means
It is the underscore sized to fit East Asian text.
Chinese, Japanese and Korean characters typically fill a square cell, while Latin letters and marks are narrow. Japanese character sets of the 1970s and 80s therefore included wide (zenkaku, “full-width”) copies of Latin letters, digits and punctuation, so they line up neatly with kana and kanji. Unicode kept those copies for compatibility, and _ is the wide twin of the basic underscore _. In Chinese typesetting it also has a job of its own: as the proper-noun mark, a line drawn under names of people, places and institutions.
To a person, both mean the same line. To a computer they are different characters: an exact search for _ may miss _, and systems that accept only ASCII, the basic Latin character set, refuse _.
Why did I get _ instead of _?
Your keyboard was probably in Japanese (or Chinese) input mode.
In many Japanese input setups, symbols come out full-width, so the underscore key gives _. Switch to direct or half-width input, or convert the text, and you get _ again. Chinese input methods have a similar full-width mode.
Neutral punctuation; in kaomoji it becomes a mouth or a flat line of a face.
In a username, password, email or code, _ is a different character from _ and may be rejected or fail to match.
Part of the collection
Don't confuse it with
2Combination ideas
4Questions people ask
2Why won't a website accept my username with _?
It sees a different character. Retype it as _ with your keyboard in English or half-width mode.
Is _ the same as _?
Same meaning, different character. Systems that apply compatibility normalization (NFKC) turn _ into _; exact comparisons keep them apart.
Collections with _
7How to type or find _
- Also called
- fullwidth underscore · wide underscore · zenkaku underscore
- Search it as
- underbar · Japanese underscore · _
- Shows up in
- Anything that takes text · its shape is whatever your font draws
_\FF3FCodes and encoding1 code points
A visible symbol can contain several code points. JS .length counts UTF-16 units; UTF-8 counts bytes.




