According to Schematron inventor Rick Jelliffe:
The Schematron differs in basic concept from other schema languages in that it not based on grammars but on finding tree patterns in the parsed document. This approach allows many kinds of structures to be represented which are inconvenient and difficult in grammar-based schema languages.
Language is based on XPath
Validator is implemented in XSLT
W3C Schemas are conservative: everything not permitted is forbidden.
Schematron is liberal: everything not forbidden is permitted.
No data typing; validation only
Handles unordered structures very well
Handles descendant constraints very well
Almost self-documenting