Moral Platonism

Moral non-naturalism is the view that there are stance-independent moral facts that are not reducible to non-moral facts. Moral platonism is a version of moral non-naturalism that explains the grounding of the moral facts, and our knowledge of necessary moral truths by means of universals (objects that can have multiple instances). My primary goal in this essay is, rather than arguing in favor of moral platonism, to simply offer an account of moral platonism that shows that criticisms of it are misguided.

Metaphysical grounding and knowledge of morality

Some opponents of moral non-naturalism question whether it can offer a coherent account of the grounding of moral facts, or our knowledge of those facts.

But there is really nothing especially mysterious about the ontology and epistemology of non-natural moral facts.

Consider instead a non-moral relation, for example being in front of something. I can recognize that my computer is in front of me. Further, I know the necessary truth that nothing that is entirely in front of something is also behind it.

Now, our knowledge that necessarily, nothing in front of something is also behind it is based on first imagining or perceiving a particular scenario, recognizing the relation of in front of, and realizing that in the particular case the relation of in front of and behind of are incompatible, and then abstracting away from the particular details to realize that it is a truth that applies in all possible cases.

Further, the fact that the computer that is in front of me is not (and cannot be) also behind me is grounded in the more general fact about universals that the relations in front of and behind of are incompatible. Some other examples of platonic grounding are where a particular four-feet stick is longer than a particular three-feet stick, due to the relation between four-feet and three-feet (which are universals); that three apples are more than two apples, due to three being greater than two; that a particular banana’s yellow is more similar to a particular apple’s red than a particular blueberry’s blue, due to yellow’s being more similar to red than blue; etc.

We can think of morality analogously. Morality involves the relation of obligation, perhaps a two-place relation between a volition and some state of affairs, or perhaps a three-place relation involving the agent and the aforementioned things.

If one admits that there is no problem of grounding or knowledge in other cases of relations, it is not clear what would be problematic about the case of morality, since we can say exactly analogous things in this case. Facts about abstract universals (such as obligation) ground particular cases of moral obligation. I am obligated, for example, not to kick the wall on a whim, because that would cause me to have a bad experience of pain, and there is a fact about the relevant universals (for example, obligation, badness or suffering, volition) such that suffering necessarily relates to volition in a certain way, namely that of being negatively obligating. Further, I can come to know more general moral facts by abstracting from particular cases. From considering how I should not kick the wall because of the perception of pain that it would cause, and perhaps considering some other cases, I can come to see that in general, I should not cause myself suffering. And I can further abstract away from the prudential cases to see that I should not cause others suffering either.

Some proponents of theistic metaethical accounts object to non-theistic accounts for being unable to ground morality, but if one admits that God is unneeded to explain other cases of properties and relations, it is unclear why God would be required to explain grounding in this case. Anyway, what I said above seems like a perfectly coherent account of how the grounding of particular moral obligations works.

Now, one might object that there is further need of grounding to explain not only why, for example, I should not kick the wall on a whim, but even facts such as that suffering is bad. But either there are eventually going to be some brute facts, or there is an infinite regress, or there is some self-grounding fact. Classical theists might insist on the last option. Of course, at this point, the objection would not be specific to morality. Anyway, it is not clear to me if it is coherent to talk about a self-grounding fact, and if it is, I do not see how it could be God (if God is supposed to be a person).

Besides grounding concerns, some might doubt whether we can abstract away from a particular situation to gain universal knowledge, or question whether universals exist. But this suggests that there is no grounding or synthetic a priori knowledge at all, not just in the case of morality. At any rate, even the nominalist could be a moral non-naturalist, but simply say that all cases of moral obligations are brute facts and we cannot know a priori what our obligations will be (though maybe we can guess based on induction?). While I don’t think that that is a plausible position, it would be no less plausible than some versions of moral anti-realism that would also need to take such a moral particularist view – the moral subjectivist, for example, would need to say something similar, unless there are necessary psychological laws requiring that we have certain stances towards certain choices.

Inability to communicate moral concepts

Another objection that has been offered against moral non-naturalism is that moral non-naturalists are supposedly unable to communicate the moral concepts that they use.

Lance Bush has made such a criticism of ‘external reasons’ (I cannot be sure, but I assume that he would say something similar about a supposed relation of obligation – I think that talk of ‘reasons’ is just another way of talking about obligations):

I agree with Williams’s instincts in this passage. I think the notion of a “reason” that Bentham invokes, which as far as I can tell accords with Parfit’s notion of an “external reason,” isn’t simply mysterious or obscure, but that it has no meaning whatsoever. At least one red flag (though not a dispositive one) is that proponents of the notion of an external reason of this kind appear unable to communicate the meaning of the alleged concept. That is, they can’t explain to others, without vacuous and vicious circularity, what they even mean. Now, could it be that there are incommunicable concepts of this kind? I grant it as an epistemic possibility. But based on my current understanding of how concepts work, I think the answer is a firm no.

My response to this sort of complaint should be unsurprising. Just like in the above cases, morality is analogous to other cases. There is a sense in which the notion of obligation cannot be communicated, but this is simply the sense in which we cannot define it. While some people might naively complain about words that cannot be defined, this is not what Lance Bush is doing, who later on acknowledges ‘there may very well be “atomic” concepts that comprise “compound concepts.”’. (He understates things here, though, by saying ‘there may very well be’; there surely are concepts that cannot be explained by simpler concepts, or else there would be a vicious regress.)

Now, maybe there are some people who talk about ‘reasons’ without being clear that they are just referring to a type of relation, in which case there might be more reason to be skeptical. But if all we are talking about is a type of relation, I don’t see how there is a concern about the inability to communicate the concept of it.

In the sense of not being able to define it by using words that aren’t mere synonyms, it is true that it cannot be communicated. However, this is true in many cases. I also cannot communicate by definition the concept of red or of longer than. On the other hand, I could communicate those concepts by using examples.

Likewise, the concept of obligation can be communicated by examples. Imagine some case where there is a decision to be made. For example, the thought of punching the wall briefly flashes across your mind; in such a case, you have a choice to punch the wall on a whim or not to do so. But in such a case, there is an obligation not to do so, since punching the wall would lead to an unpleasant experience. Or instead imagine a case of a positive obligation to do something. Suppose you are almost out of your favorite condiment; you could go to the store and buy some more which would lead to the future pleasure of delicious food, or make do with boring food.

Just like you can communicate the concept of red by giving examples of red objects, or communicate the concept of longer than by offering examples of objects that are longer than other objects, you can likewise communicate the concept of obligation by offering examples of decision-making where you would have a stronger obligation to do one thing rather than another.

Now, at this point, maybe an opponent of moral platonism (or moral non-naturalism more broadly) will admit that there is no a priori reason to think that we cannot perceive moral obligations or otherwise take issue with non-natural moral obligations, but instead insist that when he considers such cases, he does not perceive any such obligation, leading him to think that people who talk of such a thing are speaking nonsense. If there is such a person, not much can be said to him. It would be like trying to give a bunch of examples of something that is in front of something else, and the person you are trying to explain it to is incapable of grasping the concept of in front of. Of course, maybe the moral non-naturalists have simply deluded themselves into thinking that there is some relation where there is not. But either way, whether the hypothetical opponent is simply incapable of grasping the concept, or whether the proponents are delusional, there is no way to settle the dispute.

However, I think that a much more plausible route for the moral anti-realist to take is to insist that what we are taking to be some stance-independent moral obligation is instead some stance-dependent relation. Maybe when making a decision, or considering an imaginary scenario, we are simply more attracted to one option over the other, and confusing an emotional reaction with something that is independent of how we feel about the potential decisions.

While my main purpose in this essay is not arguing in favor of moral platonism, but simply offering a plausible account, I think that it is worth briefly arguing against this possibility which might initially seem plausible.

First, it seems that there are cases where we are more attracted to one decision but still recognize that we ought to do something else. For example, in the case above where I mentioned going to a store to buy something rather than going without it: someone might feel intense dread at the prospect of going to the store and so feel better about going without the thing, but still recognize that he ought to simply go to the store to buy it. There are, I think, cases where one might recognize that something is what one ought to do, but not do it due to an aversion.

Further, moral subjectivism cannot account for moral reasoning. I can recognize that it is a necessary truth that, other things being equal, I should not cause myself suffering. This is perhaps even clearer in non-prudential cases. Moral arguments rely on synthetic a priori knowledge. For example, the argument from marginal cases relies on the knowledge that there must be some non-physical difference between individuals for there to be a moral difference. This sort of reasoning cannot be explained at least by a simple version of subjectivism, since there are presumably no necessary psychological laws such that we cannot have arbitrary stances. We can disapprove of torturing babies but approve of torturing non-humans that have no non-physical differences from the baby. In general, arguments in moral philosophy rely on the axiom that there must be some sort of non-moral difference for there to be a moral difference, and also rely on ruling out certain differences as being morally irrelevant (such as species, race, sex, etc.). (See also the section ‘Ability to reason about morality’ in a previous post.)

Conclusion

People express various concerns about non-theistic moral non-naturalism, such that it cannot account for grounding, or that it is somehow incoherent, but the sort of a priori concerns people give apply in other cases, and further, it seems that a coherent account of moral platonism can be offered which addresses the concerns. At the very least, one must admit that relations exist and that we can perceive them. Either there will be some account of grounding in the non-moral cases that can be applied to the moral case, or in the other cases there is no grounding, in which case we can say the same thing about morality – however, I think that platonism offers a satisfying account of the grounding of at least some moral facts (the ones that seem derivative, like that I should not stab myself right now; it is unclear if the more fundamental ones can have any grounding, since either there must be something brute, an infinite regress, or some self-grounding thing). Likewise, whatever is said about our knowledge of relations and necessary facts about them can be said in the moral case – probably that we are somehow aware of the relations, and that we can then mentally remove some of the details for more general knowledge. Lastly, there is no a priori reason to suppose that the notion of obligation is incoherent due to the inability to communicate it – in the sense that other atomic concepts can be communicated, it can be communicated, and in the sense that they cannot, it cannot.