⼠
| ⼠ | |||
| Unicode name | Kangxi Radical Scholar | ||
|---|---|---|---|
| Code point | U+2F20 | Radical | 33 |
| Block | Kangxi Radicals (U+2F00–U+2FDF) | ||
| Character | 士 (U+58EB)
| ||
| Decomposition | Compatibility, to U+58EB | ||
| Normalizes under | NFKC, NFKD | ||
⼠ is a Unicode character at U+2F20, named Kangxi Radical Scholar. It is not a Chinese character in ordinary use. If you arrived here after copying a glyph out of a radical chart or a dictionary index, the character you almost certainly want is 士 at U+58EB.
The two look identical in most fonts and mean the same thing, but they are separate code points, and software treats them as unrelated. A search for ⼠ will not find text containing 士.
1 What this code point is for✎
Unicode encodes the 214 Kangxi radicals a second time, in a block of their own, so that a radical chart can be typeset as a chart. A table of radicals is a list of abstract indexing elements rather than a passage of text, and giving those elements their own code points lets a font apply chart-specific shapes and spacing without disturbing how the same glyph is drawn in running prose.
The block also serves round-trip compatibility with older East Asian character sets that had made the same separation.
Neither purpose involves writing words. ⼠ should not appear in article text, page titles, search queries, filenames, or any other place where the character 士 is meant.
2 Normalization behavior✎
⼠ carries a compatibility decomposition to U+58EB. That determines exactly which normalization forms will convert it:
- NFKC and NFKD convert ⼠ to 士
- NFC and NFD leave it unchanged
This is the practical trap. Compatibility decompositions are weaker than canonical ones, and most systems that normalize silently — file systems, database indexes, web forms — apply NFC rather than NFKC. So ⼠ survives in places where a canonically-decomposable character would have been folded, and two strings that look identical on screen compare as different.
The CJK Radicals Supplement block is worse in this respect. Of its 115 characters, only two carry any decomposition at all, so the remaining 113 will never normalize to anything under any form. A radical glyph copied from that block is permanently distinct from the character it depicts.
3 See also✎
4 Further reading✎
- The Unicode Standard, Chapter 18: East Asia — defines the Kangxi Radicals and CJK Radicals Supplement blocks and their intended use
- Unicode Standard Annex #38 — the Unihan database, including the radical-stroke indexing this block supports