Compatibility form of 梨

The same thing as 梨 — “pear” or “pear tree.”

梨 is the everyday Chinese character for the pear, read nashi in Japanese and used as hanja — the Korean name for Chinese characters — in Korean. It is built from 木， “tree”, which carries the sense, and 利， which once supplied the sound — today the clue reads lì while the pear stays lí. 梨 is that same character under a second code point: a character's numbered slot in the Unicode table.

The character is genuinely old: the Shuowen Jiezi, a second-century dictionary, already defines it as a fruit, and it turns up in the classical texts Zhuangzi and Huainanzi. Ancient scribes also borrowed it for other words — “to cut”, “the multitude” — but the pear meaning is the one that counts today.

Why does a second code point for 梨 exist?

So old Korean-encoded documents can round-trip through Unicode intact.

Korea's industrial character standard KS X 1001 listed this pear hanja at its own slot, separate from the unified form. When Unicode merged Chinese, Japanese and Korean characters in the 1990s it kept one code point for each such legacy entry — this one's source is K0-6C31 — so files could convert to Unicode and back without losing which slot each character came from.

Hundreds of these preserved duplicates sit in the same block. Despite the “compatibility” name, the equivalence here is canonical — the strict kind — so every Unicode normalization form, even plain NFC, rewrites 梨 as 梨. Whether a given program normalizes text is another matter; when it does, the distinction is gone.

Which one should I write?

Write 梨； keep 梨 for legacy data that must be reproduced exactly.

Chinese, Japanese and Korean input methods produce the ordinary 梨； this form survives in archives, migrated databases and edge cases where the original encoding must be reproduced exactly. A font may draw the two identically, draw 梨 in a different style, or have no glyph for it at all.

Because comparison after normalization treats the two as one character, the choice rarely changes what a reader sees. It does change the underlying code point — usually the very thing being checked.