Moral realism is the doctrine that some propositions asserting that some action is ‘morally’ good (obligatory, bad, or wrong) are true. This paper examines three different definitions of what it is for an action to be ‘morally’ good (with corresponding definitions for ‘morally’ obligatory, bad, or wrong) which would make moral realism a clear and plausible view. The first defines ‘morally good as ‘overall important to do’; and the second defines it as ‘overall important to do for universalizable reasons’. The paper argues that neither of these definitions is adequate; and it develops the view of Cuneo and Shafer-Landau that we need a definition which is partly in terms of paradigm examples of morally good actions, which they call ‘moral fixed points’. Hence the third and final definition is that an action is morally good if it is ‘overall important to do because this follows from a fundamental universalizable principle, belonging to a system of such principles which includes almost all the moral fixed points; when a suggested fundamental principle is one which would be shown to be very probably true by the exercise of reflective equilibrium over many centuries’.
