Metaethics

3. Metaphysics

The metaphysical branch of metaethics, or moral metaphysics, concerns the nature and existence of moral facts, properties (like goodness), relations (like being a reason for), and entities (like obligations) (see Value, ontological status of). The broadest theoretical distinction here is between moral realism, which accepts the existence of some such things, and moral antirealism (or irrealism), which rejects it (see Moral realism). Philosophers who are noncognitivists about moral language are invariably anitrealists, but not all philosophers who are cognitivists about moral language are realists. One important form of moral antirealism is error theory (e.g. Mackie 1977), which maintains that moral language refers to supposed moral facts and properties that don’t really exist. Error theory can be understood by analogy to atheism, which maintains that theological language refers to supposed divine entities like God that don’t really exist.

A central metaphysical puzzle for realism concerns what the nature of moral properties, etc., could be. Realist views are commonly divided into two camps, ‘naturalism’ and ‘nonnaturalism’. The nonnaturalist position is most frequently associated with Plato and Moore. It maintains that moral properties are categorically different from the kinds of natural properties that we encounter through our senses and investigate through empirical science. The postulation of nonnatural properties beyond the ken of science has struck many philosophers, however, as intolerably bizarre or ‘queer’ (in J. L. Mackie’s (1977) influential vocabulary), although nonnaturalists pursue a variety of strategies to argue that this response is mistaken.

Metaethical naturalists by contrast maintain that moral properties are just ‘natural’ properties like any other, and therefore not at all bizarre. The challenge for this type of realism is to give a plausible account of exactly which natural properties these are, as it has seemed to many philosophers that any moral properties would have to have special features that are not to be found anywhere in the natural world. To identify moral properties with natural properties, Moore influentially claimed, is therefore to commit a ‘naturalistic fallacy’ (see Moore, G.E. §1). We discuss some of these features below, in the sections on epistemology and psychology (§§4–5). A characteristic motivation for antirealists of both the noncognitivist and error theorist varieties is the view that while any moral properties would have to be nonnatural, the idea of nonnatural properties is so bizarre that it beggars belief. (One family of views about moral metaphysics that does not fit neatly into either the nonnaturalist or the naturalist camp is metaethical supernaturalism, like the view that to be morally right is to be commanded by God. While it may seem appropriate to categorize supernaturalism as a form of nonnaturalism, nonnaturalists generally maintain, following Moore, that the reasons not to identify moral properties with natural properties are equally reasons not to identify them with supernatural properties. Moral properties, they think, belong in a category all of their own.)

One metaphysical puzzle about morality, which many have thought presents a special problem for non-naturalism, is the so-called supervenience of the moral on the nonmoral (see Supervenience). Metaethicists are in almost unanimous agreement that moral properties ‘supervene’ on natural or nonmoral properties, which is to say that there can be no difference between two possible situations in their moral properties unless there is also some difference in their natural or nonmoral properties. For example, it cannot be the case that one person is evil while another is not, if there is no difference between them that can be described in nonmoral (particularly psychological) terms. Nonnaturalism seems to have a special difficulty in accounting for this relationship, since it maintains that moral properties are distinct from natural properties.

Some metaethicists suggest that the moral realism debate as described above is wrong-footed, because it focuses on trying to find referents for so-called thin moral words like ‘good’ and ‘right’. They point out that this overlooks the richness of our moral vocabulary, which includes so-called thick words like ‘brave’, ‘honest’, and ‘cruel’. It does not seem nearly as tempting to be sceptical about the existence of any real properties in the world that are the referents of these terms.