Good and bad news in formalizing generalized implicatures