What Metaethics Asks

Two people argue about whether it was wrong to break the promise. A third says that nothing is ever really wrong, and that both of them are expressing preferences in a grammar that disguises them as reports.

The third person has not joined the argument. They have said something about the argument, and about what the two participants were doing when they had it. Sorting out which question is which is the last piece of vocabulary this tier supplies.

Two levels

Everything in this tier so far has been normative ethics or preparation for it. The theories of the fourth article are answers to which considerations make actions right. The trolley pair tests those answers. The method of the sixth article is how such answers are supported.

Metaethics steps back from all of that. It takes the practice of arguing about right and wrong as its subject and asks what the practice involves. A metaethicist can hold any first-order view or none, and two people who agree completely about which actions are wrong can disagree completely at this level.

Notice that stepping back is not neutrality. Saying that moral claims are expressions of attitude rather than reports of fact is a substantive and contestable claim, not a refusal to take sides.

The first question: what are we saying?

"Torturing children for entertainment is wrong" looks grammatically like "arsenic is poisonous". Both are declarative, both take a subject and a predicate, both can be embedded in conditionals and questioned and denied.

Whether the resemblance is real is the question. On one family of views the sentence states something that is straightforwardly true or false, and morality is a subject matter about which people have beliefs. On a rival family the surface grammar is misleading, and the sentence functions to express an attitude of disapproval or to prescribe rather than to describe. The intermediate tier calls these families cognitivism and non-cognitivism and works through the main versions of each, along with the hardest objection to the second, which is that it struggles to explain what moral sentences are doing when they appear inside conditionals and arguments.

The second question: is there anything there?

Grant for the sake of argument that moral claims do state facts. Then the fact that torture is wrong is a fact about something, and an account is owed of what wrongness is and where it fits.

The options in play range widely. Some hold that moral properties are natural properties, identical with or constituted by facts about wellbeing or harm, and therefore no more mysterious than facts about health. Some hold that they are real and irreducible, a distinct kind of feature that the natural sciences do not describe. Some hold that moral claims do state facts and that all the positive ones are false, since the world contains nothing of the kind they are about, a position J. L. Mackie defended in 1977 partly on the grounds that objective values would have to be entities of a very strange sort. And some hold that moral facts exist but are constituted by the output of a procedure of reasoning rather than found. The intermediate and advanced tiers take each of these in turn.

The third question: how could anyone tell?

The previous article was already doing this, without the label. Asking whether an intuition is evidence is asking a question in moral epistemology, and the method of reflective equilibrium is a proposed answer to it.

The area is harder than its counterpart elsewhere because the obvious source is missing. Claims about the natural world are checked against observation, and there is no moral observation. Nobody sees the wrongness of an act in the way they see its speed. What is left is intuition, inference from other moral claims, and inference from non-moral facts, and each has been challenged. The persistence of moral disagreement is standardly raised here as evidence that whatever the sources are, they are not working well, and the intermediate tier examines whether that argument succeeds.

A fourth strand

Running through the three is a set of questions about the relation between moral judgement and motivation. Someone who sincerely judges that they ought to keep a promise seems to be at least somewhat motivated to keep it, and a person who says they judge it and feels nothing invites the suspicion that they do not really judge it.

If that connection is a necessary one, it constrains everything above. A view on which moral judgements are beliefs about ordinary facts owes an explanation of why this particular belief comes with motivation attached, since beliefs about the ordinary world generally do not. A view on which they are expressions of attitude gets the connection for nothing and pays for it elsewhere. The intermediate tier treats this under the Humean theory of motivation, and it is where the semantic and metaphysical questions turn out to have teeth.

What this tier has done

Seven articles have supplied vocabulary rather than conclusions, which is what a beginner tier owes.

Claims about what is were separated from claims about what ought to be, and the ought was analysed into contributory reasons and all-things-considered verdicts. A moral theory was distinguished from a list of verdicts, and from the decision procedure it may or may not recommend. Three families of theory were set against a single case and shown to disagree about the question before they disagree about the answer. A designed pair of cases exposed the distinctions that deontological theories rest on, and the difficulty of stating a principle that yields both verdicts. The method that all of this relies on was named and its main objection stated. And the questions about morality were separated from the questions within it.

The intermediate tier takes the theories one at a time with their developed objections, works through the metaethical positions named here, and takes up motivation and reasons, which is where the two levels meet.