The
P5831 "usage example" has a constraint (as you can see broken) for only using on the Lexeme main statement, and my way of loosening that constraint actually coincides with the Lexeme Data Model even more...where Senses are where we are allowed to place Synonyms, Translations, etc. to capture the nuances of a Lexeme with many Senses and not limited to only 1 Sense. Allowing
P5831 "usage example" at a Sense level to accurately portray the Sense (true meaning). Loosening of the that constraint will be especially important for generic words with 10's to 100's of Senses, and then further see similar usages, even across synonyms. This helps build multilingual knowledge more, not constrained to less. For example:
car (en) versus
carriage (en-gb)