The moral pragmatist’s wager

I don’t write about meta-ethical epistemology often enough. Time to remedy that.

Something occurred to me as I was reading Jonathan Jong’s blog a few weeks ago. Jonathan said: “Most people want to be moral realists, to believe that there are such things as moral facts beyond our social conventions.”

I think that’s pretty much right. A lot of people do believe that there really is, objectively, right and wrong, and that this isn’t just a matter of following social conventions – and nor is morality merely a matter of conforming to our biological nature, nor satisfying personal preferences, nor complying with a set of laws, nor (as many would argue, given the preceding two points) obeying divine commands.

For morality to truly be morality, it would have to be something quite distinct from all these things – although that’s not to deny that conventions, preferences, scriptures and the rest can shape our thinking about what’s right and wrong. As well as being a notable school of thought in ethical theory (‘non-naturalism’), this view does in rough outline have appeal for a lot of people whose heads haven’t been saturated with academic philosophy.

And yet… if morality is to be something wholly standalone, then what exactly is it? What sort of things are moral principles? And, more practically, how can we come to grasp them? It seems that empirical observation and logical reasoning aren’t up to the job of delivering knowledge of these peculiar facts, so are we left to rely on some special sort of intuition? But that, too, is an obscure idea.

We want to be moral realists – because we want to be able to be moral people – but we’re not at all sure how to justify belief in objective moral truths (although this doesn’t prove that there are no such truths). So, are we stuck with a puzzled agnosticism that will paralyse us when we face up to it or awkwardly nag at us when we try to ignore it?

I think we may be able to do a bit better.

Pascal’s wager

In the same blog post, Jonathan also wrote, a few sentences earlier: “Perhaps there are good prudential or pragmatic reasons for assuming that God does exist, even if this assumption is not ‘belief’ in the proper sense. It's less, ‘I hold it to be true that God exists’ than, ‘I live as though it is true.’ Perhaps.”

And, although he was talking about Kant, this put me in mind of that more famous pragmatic religious argument, Pascal’s wager. This is not an attempt to prove god’s existence but a cost-benefit analysis to recommend the rationality of believing in god. It’s aimed at people who tend to like the idea of god but just don’t find themselves convinced that he exists.

Briefly: you can either believe in god or not. If you don’t believe in god and he doesn’t exist, then you lead an ordinary finite life. But if you don’t believe in him and he does exist, then you go to hell for eternity, which is the most awful thing that could happen to you. Your other option is to believe in god: if you do, and he doesn’t exist, then you’ve wasted some time and effort and perhaps money in your involvement with church activities, which is a loss, but only a modest one (you may also gain a sense of fulfilment, but the wager can assume that this is outweighed by the losses here). But if you believe in him and he does exist, then you’ve hit the infinite jackpot of heaven.

So the outcomes for you if you don’t believe are either unremarkable or utterly disastrous; if you do believe, your outcomes are either modest inconvenience or eternal bliss. Assuming you don’t rule out god’s existence completely, then even if you judge there to be only a slim chance of it, the average expected outcome of belief is immeasurably better than that of disbelief. So it’s in your interests to believe.

There are plenty of objections to Pascal’s wager. Here are some:

1. It assumes voluntarism about beliefs. But we can’t simply decide to believe in the truth of something by an act of conscious will. How many emails did I receive yesterday? You don’t know, but try, now, sincerely to believe that I got 15. That may sound like a plausible number, but you really have no information. Can you firmly believe it, rather than merely entertaining the idea? No. You could freely choose to act as if you believed this, but belief itself isn’t like that. Something has to convince you.


2. Even setting the above aside, it’s an abdication of rational judgement. While it does trade on rationality, this is of a calculating self-interested kind rather than the epistemic kind that judges propositions to be true, false or uncertain based on the available evidence. If the logic of the wager were extended across other walks of life, we would find that wishful thinking got in the way of our ability to judge true from false.


3. It subverts whatever virtue there may be in belief in god. We might accept that believing in god is something that he would esteem as morally praiseworthy, but given the psychology of the wager – self-interested calculation to maximise personal advantage – it’s hard to see any virtue in this at all.


4. It assumes the only relevant options are this god in particular (Pascal was a Catholic) or not. But the same argument could be made about the benefits of adhering to a number of different religious views. Accepting one means rejecting all the others, and if it’s equally rational to adopt belief in any one, then the wager’s logic could push you in many contradictory directions with equal force. Which means that it can’t justifiably push you towards any of them.

What does such a bad argument about religion have to do with wanting to believe in objective morality but not being sure how to get a handle on the idea? I think the structure of Pascal’s wager can be adapted, in a way that avoids equivalent objections, for people who’d like to be able to ‘do the right thing’ but don’t know how to justify the view that there really is a ‘right thing’ to do.

Betting on moral pragmatism

You can live you life either on the basis that there are moral truths (beyond conventions, preferences, etc.) or not on that basis. (a) If you disregard morality and indeed there is none, then you live in a way consistent with your non-moral desires, many of which you may satisfy, and it doesn’t morally matter. (b) But if you live in such a way and there is an objective morality, then while satisfying those personal desires, you will likely lead a morally bad life.

Your other option is to live with the working supposition that there are objective moral truths. (c) If you do, and there aren’t, then you sacrifice some of your personal desires and achieve nothing of moral value (although you may gain a sense of moral fulfilment, but let’s assume that this is outweighed by the losses here). (d) But if you live in such a way and there are moral truths, then you can adhere to these and achieve some good.

So the outcomes for you if you don’t live as though there are objective moral truths are a (potentially) materially satisfying life that is either morally neutral or positively immoral. If you do live as though moral realism is true, then your outcomes are either inconvenience or moral goodness. These different types of factor are harder to weigh than the purely personal costs and benefits of Pascal’s wager.

If we could assume moral realism, then it would be easy: moral considerations trump material ones. From the initial doubtful position, all we can say is that if there are any moral considerations, then they trump material ones. Two of the four outcomes do specify moral realism, so these will count as (b) very bad and (d) very good. The nihilistic other outcomes are morally neutral and either (a) materially satisfying or (c) materially unsatisfying.

Someone who’d like to be a moral realist and is considering this wager would agree that morality, if it is real, is by definition more important than personal advantage. This then requires the judgement that the moral badness and goodness of (b) and (d) respectively is greater not just in degree but in kind than the material goodness and badness of (a) and (c). Which means that acting as if there are moral truths leads to the best expected average outcome, as long as you don’t rule out moral realism completely.

So you may not be able to prove that morality is real, beyond conventions and preferences and the like, but the only way to have even the possibility of adhering to it is to act as if it is real. That’s the moral pragmatist’s wager.

Do similar objections to those that scupper Pascal apply here? Let’s go through them:

1. Voluntarism about beliefs isn’t relevant. The moral pragmatist’s wager isn’t about choosing to believe in moral realism even though you’re unconvinced; it’s about choosing to act as if moral realism is true.


2. Because it’s not about belief, it can’t be an abdication of rational judgement. But might it be an abdication of moral judgement, given your uncertainty about whether the moral principles you act on are really real? No. Because even if you’re uncertain about whether there are objective moral truths, acting as though there are is the only way you even stand a chance of doing the right thing and being a moral person. The driving motive in taking the wager is that you want these moral goods, and choosing to go the other way would guarantee that you won’t get them.


3. It can’t subvert whatever virtue there may be in belief in moral realism, because it doesn’t mandate such belief. But perhaps acting on the basis of a moral belief that you don’t actually hold might somehow negate whatever moral value such actions have? It’s hard to see how this might be, though. To the extent that morality involves consequences of actions, these are unaffected; to the extent that it involves intentions or personal virtues, the wager is based on the intention to be as moral as possible, and this is liable to cultivate a virtuous character; to the extent that it involves rules or duties, these can be adhered to equally well on a pragmatic basis. Also, given that our beliefs aren’t voluntary, there can’t be anything reprehensible about not fully believing in something of which you’re genuinely unsure.


4. Does it assumes the only relevant options are morality or none? What about different moral viewpoints? It’s true that there are such varied views, which might seem to pose a problem. But, as I said in the very first sentence, this argument is about meta-ethics. It’s about the very general nature of moral truth and moral knowledge at all. The fact that it doesn’t tell us which particular moral theory to favour, or how to resolve any given dilemma, is no criticism. The moral pragmatist’s wager simply tells us that, if moral realism is appealing but not proven, it can still make sense to treat it as a working assumption and to keep thinking morally about more specific issues.

So, there it is. Having only thought of this very recently, I can hardly claim that this line of thinking is what shapes my moral outlook. But it does seem to me (and I add the caveat that this isn’t my field) that the argument more or less hangs together. Maybe not, though; maybe it’s an old idea that’s long been dismissed. Maybe this is just an amateur’s light mental workout, or maybe it’s a striking new piece of moral philosophy.

Wanna bet?