George Saunders on REVISION and CAUSALITY