Of Metaethics, Error Theory and What Morality Really Is

I’ll say it again: doing metaethics is a dreadful way to spend one’s time. Yet, here I am. Doing metaethics. For, like doing tax returns and scrubbing the bathroom, there are some unsavoury endeavours that are necessitated by our chosen course of life. And as my chosen course involves walking the paths of ethical theory, I’m forced to wade through the swamps of metaethics from time to time. So, don your galoshes and on with the show.

I stated recently that Sam Harris ought to be a moral anti-realist, and in shifting to such a stance, he’d lose little and gain much. Namely, he’d lose the mad-dog moral naturalist realism that insists that science can determine human values – and in doing so, evaporate the ire of the manifold philosophers who’ve criticised this aspect of his approach.

What he’d gain is an ability to talk about moral facts, or facts that pertain to making a moral judgement. This has got me into some metaethical strife, according to Richard Wein. Why? Because I’m getting all error theorist on Harris’ realism, yet I’m still talking about moral facts. But, if error theory and anti-realism are to be taken seriously, then moral statements are all false. The only point in continuing to talk about them as if they’re real is to pretend they’re real, a la moral fictionalism.

Let me elaborate. And brace yourself, this is going to get metaethical.

Moral realism suggests there are facts about morality, and they exist in some mind-independent manner. They’re woven into the world. This means (according to a grossly simplistic reading) we need only discover what these moral facts are, and we know how we ought to behave.

Moral naturalism suggests these facts about morality are natural facts, and Harris’ brand believes science can reveal these facts. Non-naturalism, on the other hand, believes moral facts are not concomitant with natural facts, but let’s ignore non-naturalism for the time being, mainly because it’s bonkers.

Moral anti-realism sits in opposition to moral realism, and says moral facts either don’t exist or, if they do, they do so in a mind-dependent manner, such as in a subjectivist or constructivist reading.

Then we have cognitivism, which reckons moral utterances – things like ‘harming innocents is wrong’ – are statements of fact, and are the kinds of things that are true or false. Non-cognitivism says the opposite: moral utterances are not ‘truth-apt’ but are expressions of our attitudes and/or demands that others conform with our attitudes.

Now we come to error theory. This says there are no objectively true moral facts, because such facts would be quite bizarre. Not only would these facts have a strange relationship with other natural facts, but knowledge of these moral facts would compel behaviour in some oddly binding way. If they were to exist, they would be quite queer things.

Error theory also says we must take a cognitivist approach to moral utterances – which leads to Mackie’s famous conclusion:

I conclude, then, that ordinary moral judgements include a claim to objectivity, an assumption that there are objective values in just the sense in which I am concerned to deny this.

Error theory about morality is like error theory about phlogiston. When people used to prattle on about this substance that was responsible for combustion, and called it phlogiston, they were talking about some thing they presumed to exist. But it didn’t. They were in error, and their discourse was tainted.

As a wee adjunct to error theory, we have moral fictionalism, advocated primarily by Richard Joyce (2001). This accepts error theory as true, but says that moral discourse is useful, so we can continue talking about moral facts but accept that we’re talking with our fingers crossed behind our backs while winking suggestively.

The implication of error theory is that statements like ‘harming innocents is wrong’ is just false. There might be natural facts about harm, and about whether Alice is innocent or not, but the moral judgement itself is just false. It’s like saying ‘that piece of burning wood is releasing phlogiston’. There might be facts about the wood, and whether it’s burning, but the statement about phlogiston is just false. Right, moving on.

Now, this is a traditional account of error theory, and I’m largely sympathetic with it. But I don’t think it’s the end of the story. Error theory is a negative thesis, and a useful one. However, it only targets one particular rendering of morality, namely the moral realist version. But the moral realist rendering of morality as being objective, binding, factual etc is actually quite an extreme account of morality itself. It’s only an historical accident that Western philosophy has become so obsessed with this impression of morality, probably because of the dual influences of Plato, and his Form of the Good, and Christianity, with its insistence that morality is objective and justified by a force external to ourselves, i.e. God.

This view of morality is wrong, and I think error theory does a splendid job of chopping it down. But error theory is like a sceptical theory targeting a particular account of knowledge. Or an atheistic approach to God. It knocks down that theory, but it doesn’t necessarily preclude different accounts of what morality is. It engenders moral scepticism against this brand of moral realism, but it doesn’t imply moral scepticism against all brands of morality.

And morality can be rendered quite differently. Error theory might deal a lethal blow to the notion that morality is objective, factual, mind-independent etc, but it doesn’t rule out a softer account that sees morality as a system for regulating social behaviour, a system of prudential oughts and hypothetical imperatives, norms as strategies, not as binding rules. To take from Joshua Greene, there are two senses in which we talk about morality:

moral₁: of or relating to the facts concerning right and wrong, etc.

moral₂: of or relating to serving (or refraining from undermining) the interests of others.

Error theory targets moral₁, but leaves moral₂ unscathed, as long as moral₁ isn’t necessary for moral₂. And I don’t think it is. Moral₂ can stem from a collective social contract agreement, and once people have agreed to buy in to this contract, you can derive a bunch of prudential oughts from it. While we often talk about these oughts erroneously as moral₁ statements, in actual fact they’re moral₂ statements. Greene even recommends we dispense with moral₁ statements altogether, but not moral₂ statements.

Sure, there’s a lot of work to be done to flesh out the nature of this agreement, how it works, how moral judgements motivate behaviour, and so on – I’m doing some of that in my thesis. But the point is that moral₂ morality looks very much like moral₁, except without the dodgy metaphysical baggage of objective moral facts.

So when I say “I’m an anti-realist about morality because I don’t think there are moral facts, a la Mackie. But I do think there are natural facts that are crucial in forming a moral judgement”, I mean I’m sceptical about moral₁, but there are facts that are relevant to coming to moral₂ judgements.

To draw another analogy, I can be sceptical about the possibility of justified true belief, but that doesn’t mean I blanket accept that knowledge is impossible. Instead I can adopt a different approach to knowledge, and get on with my life.

I think Russell Blackford would agree. He recently wrote this in his response to Harris on ABC Religion:

The moral scepticism that I favour undermines traditional moral systems that base morality in the commands of a god or in an idea of absolute moral good and evil, but it also undermines any idea that we are objectively required to do something as cosmic – and remote from our practical goals – as maximizing the well-being of all conscious creatures.

We can get by with more modest aims, such as each doing what we can, consistent with our other projects, to reduce the world’s burden of suffering.

Ultimately, it’s it the legacy of Western philosophy to conjure up these strict definitional dichotomies – existence/non-existence, moral realism/moral nihilism, knowledge/scepticism, realism/nominalism etc – and then bumble around in circles for decades trying to decide which is less wrong rather than attempting to extract itself from its own miasma of over-abstraction and reification and taking a softer, more pragmatic approach.

Bollocks to most metaethics (and metaphysics, while I’m on a roll). I agree with Harris that metaphysics is largely a barrier to producing positive moral theories. But I disagree with him in that it is (unfortunately) important to understand metaethics, if only to avoid falling into one of the already deep metaethical holes. Metaethics is rather better as a cautionary lesson in what to avoid – namely, the extremes of any particular abstract definition of morality.

For example, it seems to me that moral statements contain both a cognitivist and non-cognitivist component: our immediate moral intuitions, largely fuelled by emotion, tend to be non-cognitivist; but our moral rationalisations that are abstracted from that intuition into a general principle seem pretty reliably cognitivist. Woot. Good to know, but this makes a lot of the one-or-the-other metaethical blustering a bit pointless.

Anyway, hopefully this makes my metaethical stance a bit more intelligible. In short, I’m sympathetic with error theory because I think moral realism is bananas. I’m an anti-realist because I think morality stems from an agreement between individuals (explicit or implicit) rather than from binding moral facts. But these say little about my positive moral theory.