I think about this footnote from @martin's book at least once every couple of months
@lindsey Very delayed reply, sorry. For the 2nd edition of the book I exchanged emails with Theo Härder (one of the two authors of that paper), and he disputes that "C" was "tossed in". To him, it was always about the DBMS enforcing invariants that have been explicitly declared as part of the database schema. I've therefore revised this bit in the 2nd edition
@martin Thank you for always pursuing the truth!