On the validity of utility statements: standard theory versus Duesenberry’s