Chapter 6 showed some of the ways in which Kant’s solution to the problem of relevant descriptions produces difficulties for his theory of right. Where an agent does not match his maxim to the situation or his act to his maxim, where he acts on an erroneous means/ends judgment, and where he faces conflicting grounds of obligation, it seems that Kant’s theory of right might give either unacceptable guidance or none at all. By contrast his theory of moral worth did not get into such difficulties.

These conclusions are surprising and, at first glance, disappointing. Other universalisability tests have been proposed primarily as necessary or necessary and sufficient conditions on principles of right. Though Hare speaks of his test as proposing a necessary condition on all moral judgments, he has in mind primarily ‘ought judgments’; Singer explicitly restricts his discussion to principles of right and wrong. A theory of right has come to seem to many writers on ethics the basic and indispensable kernel of an ethical theory, to which theories of moral worth or of supererogation or of virtue are always ancillary. If Kant lacks a theory of right he has at most fragments of an ethical theory to offer.