How to regex match words? (in multiple languages)

Question

I have a simple regex search which matches [A-Za-z], however I would like it to match other characters which would be used in dictionary words for non-english languages. e.g. ü, Ø, ɵ 人参, while excluding:

• White space.
• Punctuation.
• Numbers.
• Other non-language symbols arrows, characters for drawing diagrams ... etc.

How can Emacs regex match against unicode words used in language?

I tried [:word:] but this matches numbers.

Answer 1

Use the character class [:alpha:]

E.g. [[:alpha:]]+ to match one-or-more such characters.

C-hig (elisp)Char Classes says:

[:alpha:]

This matches any letter. For multibyte characters, it matches characters whose Unicode general-category property (*note Character Properties::) indicates they are alphabetic characters.

Whereas you were using:

[:word:]

This matches any character that has word syntax (*note Syntax Class Table::).