Does God Save Moral Knowledge?

How theism may lead to moral skepticism

I believe Plato showed pretty convincingly that God makes a terrible candidate for the ontological ground of morality. Still, a small bastion of intellectual elites (myself included) think that theism might yet have an important role to play in metaethics, given moral realism—namely on the epistemological front. The idea is that without God (or other spooOOOoky metaphysics), we would have no way of knowing what the moral facts are, since they don’t really do much down here on earth. While this seems like a promising out for the prospective moral realist, I have some reservations about how well this solution actually works, and suspect it might even work in the other direction.

For my purposes I’ll assume that theism entails moral realism. There are certainly many possible theories that are recognizably theist-ish which don’t entail realism, seeing as theism is just a vague blob of theories, but we’ll just define those out of existence. I’m also not interested here about whether theism or moral realism are actually true (sorry Lance), we’re just considering how the two relate.

To lay out my worries, we’ll be framing this around the connecting premise of the moral knowledge argument (which I absolutely DEMOLISH here). This premise is essentially: If theism is false, then we cannot have moral knowledge. The idea here is that there are all sorts of debunking arguments for moral knowledge, showing that absent some sort of extraordinary metaphysics (such as theism, or something like it), any justification for our moral beliefs is undercut.

A Direct Connection

Let’s then first try and see what happens if we assume this premise to be false (and let’s also assume that moral realism is true). This would obviously mean that we can have moral knowledge without theism (congratulations on the 3rd grade reading comprehension)—there is no need for God to mediate moral knowledge.

In this case the epistemic relation between morality and theism is somewhat simple. Basically, we can know that moral realism is true. We then conditionalize on this, and theism gets a small boost. On top of this, since God would act in accordance with the moral facts (ex hypothesi), theism makes predictions about how God would act, and we can update on these.

As an analogy, you can imagine two alternate hypotheses: Presidentialism and non-presidentialism (yes, I’m pandering to my American readers). Presidentialism says that in the white house there is a president who decides how to run the US by following a document called the Constitution. On non-presidentialism there is no such president, and the country is basically run bottom up, and there might or might not be a Constitution in the white house, but in any case the country is run independent of it. Suppose now that you find out that there really is a Constitution, and get a pretty good idea of what it says. Well, firstly that’s some initial evidence for presidentialism, seeing as it entails there being a Constitution, and non-presidentialism leaves some room for no Constitution. On top of that, since I have some idea about what the Constitution says, I can begin to make some predictions about what will happen to the country if there is a president, and update for/against presidentialism depending on what happens. All is well and good in epistemology-land.

God as a Mediator

Trouble really starts to stir in the epistemic paradise once we sever the connection between ourselves and moral facts, making God the mediating link. This is what is required for the premise in the moral knowledge argument to be true. If our moral knowledge requires theism, that’s because the moral facts aren’t knowable by themselves, and God somehow must arrange the world such that your moral beliefs are accurate to some degree for you to have knowledge.

It’s now very unclear how I can know either part of the equation, i.e. God and the moral facts. I mean, it seems sort of circular: To know the moral facts I need God, but to know whether God exists, I need the moral facts to predict what he will do. So the two need to presuppose each other to justify each other. This very baseline intuition, while ultimately not correct, does point to something problematic.

Firstly, let’s see how it’s not correct. Basically, as I note here, theism isn’t just a single theory, but a catch-all for a bunch of different specific hypotheses. Just like theism might be precisified in terms of what powers God has, or what knowledge he has, it might be precisified in terms of what the content of moral realism is. So theism-1 may say that God is such and such, and the moral facts are this and that, while theism-2 fills out alternative details. Each of these hypotheses will make precise predictions about the world. Thus we can get evidence for and against theism, even if what the moral facts are isn’t fixed beforehand. As an extreme example, if a godly figure descends from the heavens (fill in details to make the divine nature of this thing convincing) and tells you that the moral facts are X, then that’s certainly some strong evidence for theism-plus-the-moral-facts-are-X, even if other hypotheses might also accommodate this occurrence.

Nevertheless, there seems to be a sense in which the initial intuition is correct. This is because we’re no longer able to trust the regular methods in moral epistemology (basically moral intuitions). By stipulation, our moral intuitions aren’t themselves sufficient to justify our moral beliefs. But even if moral realism is true, theism doesn’t necessarily secure our moral intuitions either. After all, God simply acts in accordance with the moral facts, and it’s far from certain that us having moral knowledge is part of what’s morally good. For example, if the moral facts simply were that it’s good when people have false moral beliefs, then God would presumably create a whole bunch of people with false moral beliefs. And seeing as I don’t know many who accept this theory of morality, we would be such people.

What we need to be able to trust our moral intuitions is that the most plausible theory has moral facts such that God would (likely) create us with reliable moral intuitions. Our moral intuitions do point to moral facts where moral knowledge is plausibly valuable (even if instrumentally), so that picture is at least not inconsistent or in a lot of tension. Still, there are also plenty of alternative sets of moral facts that would involve radical skepticism about morality.

To bring back presidentialism: Imagine that you know that there is a president, and thus that there is a constitution. The only way you know the content of the constitution, though, is that the president enforces “constitution class” in every 3rd grade, where students are told the content. Suppose now that your textbook tells you that the constitution says X, where X includes (or at least suggests) that your textbook would say this. This might seem to settle that the constitution says X. But if the constitution says Y (say that the country is a dictatorship, but citizens mustn’t know), that would still predict that your textbook would say what it did, while what it teaches you would be wrong.

What you need is some reason to think that X is much more probable than Y (and other “deviant” contents of the constitution), and it’s unclear what this is supposed to be.¹ Similarly with the moral facts. Even if your moral intuitions say X, then if theism is correct, you can’t be very sure that it’s actually because the moral facts are X, or because they’re the deviant facts Y. You would need some good reason, other than your intuitions, to think that X is a more plausible set of facts than Y. I’m not sure whether we can actually give good reasons for thinking that the moral facts are what we think they are given theism, and I won’t explore that very much here.

As a general gesture though, the structure of this problem is very similar to something like external world skepticism. In the skepticism case, we can be external world realists because that’s the best explanation of the data. However I’m not as sure that this is the case for moral facts. The thing about the external world hypothesis is that it compresses a lot of highly specific predictions (that turn out true) into a much simpler theory than alternatives. With moral realism I just don’t see why the moral facts being the way we think they are predicts our moral intuitions much better than deviant facts, given theism.

This point is especially strengthened by debunking arguments. A big part of the reason why the moral knowledge premise is plausible to begin with, is that our moral intuitions are neatly explained by things that have nothing to do with their accuracy (e.g. evolution), meaning we shouldn’t trust them. Theism is then supposed to ensure that we still can trust our intuitions, despite evolution. However, that only works conditional on the moral facts being such that it’d be good for us to have accurate moral beliefs. The problem is that any set of moral facts, where our moral beliefs don’t have valence will just predict that we have the intuitions we currently have, since evolution would lead to this—even given theism. For example, if what’s good is to have a universe develop life in an uninterrupted way, then God would create a world where life develops, but he wouldn’t care about our beliefs, and so it makes perfect sense that our beliefs match the evolutionary story.²

You might just want to say that on top of God making the moral facts known, we have some immediate connection, letting us know that the moral facts aren’t deviant in some way—maybe we can presumptively trust our moral intuitions absent defeaters. However, that would simply be jumping back to the direct connection. To the degree that we can be sure that God doesn’t deviantly cause our moral beliefs, to that same degree we are justified in those moral beliefs absent God. After all, whatever mechanism through which we know the facts directly must be going around God, and so must be available on atheism. So if you say that the facts being this rather than that way is a Moorean truth or something, you’re just granting the atheist moral knowledge. If you aren’t, then it must be God who’s responsible for that moral knowledge, and deviant-God would have acted in the same way.

RIP Arguments for Theism

This is not just a problem for knowing what the moral facts are (though that might be problematic enough in itself). The thing is, theism crucially relies on the content of morality to make predictions about what God will do. I find this to be a major strength of the view, since it means that it can make a lot of predictions without filling in contingent details about God’s desires. Though If theism simultaneously ruins our moral knowledge, we are no longer able to get evidence for the view, since we can’t derive predictions from it.

Again we should be careful to notice that this isn’t strictly right. The theory “theism + the moral facts are such and such” still makes predictions, and so we can get evidence for individual instances of theism. The problem simply is that the theory “theism + the moral facts are what we think they are” might be a much smaller part of the theism probability-space than first anticipated, and what the moral facts are isn’t just settled through moral intuitions. This severely complicates and weakens any attempt at providing evidence for theism.

Usually when making arguments for (or against) theism based on what God would do (such as fine-tuning), the moral facts are held more or less in the background. So we attempt to figure out what God would do based on what our moral intuitions say, and then see how well this matches what we find (in rough terms). But if what I’ve argued is correct, this is much too simplistic. After all, 100% of our credence given theism shouldn’t be given to God acting in accordance with our moral intuitions. Rather it should also be distributed across alternative contents of morality, which would predict our intuitions being this way.

Because our original conception of theism now has to compete with all these alternatives that are not obviously worse, all our credence in any particular version of theism prior to looking at the evidence should be much lower. This in turn also leads to a (probably) much lower posterior, in theism overall, since not all of the alternatives will predict the evidence very well. And even if theism in general comes out quite probable, it’s far from obvious that the conception we started with will be the best hypothesis.

This means that on theism, figuring out the contents of morality can’t just be done in the usual ways. Rather (contra myself a few weeks ago), we also need to consult empirical data, since our moral knowledge is mediated by an entity that is supposed to make a difference to the world (by, you know, creating it and all that).

It might just be that this sort of moral skepticism is unacceptable given theism, in which case you should embrace something like option 1, or moral anti-realism. If you think that God is necessary for moral knowledge, though, you will have to accept that these radically deviant moralities are possibilities (and perhaps quite probable ones), and that the nature of God and morality might not at all fit surface-level appearances.

I think this is a complicated issue, and I’m not sure how much of what I’ve said is accurate, nor how problematic it would be if it were right. If you’ve got any thoughts on this, then I’d appreciate if you could tell me why I’m wrong (or praise my giant intellect for getting everything right).

You Might Also Like:

1. In an actual scenario (such as the real world) you would of course have much other data available (like how the rest of the country is run). I think this weakens the intuition that your knowledge of the constitution is problematic here. But if you imagine that it’s really only your textbook you have as evidence, I think the scenario begins to look much more problematic.

2. Similarly something like Paul Draper’s Aesthetic Deism would fit this. This doesn’t strictly fit the definition of theism I stipulated at the start, as it doesn’t entail anything about moral realism. But it could be fitted into this definition by stipulating that the moral facts are what Draper thinks the aesthetic facts are.