Plantinga's argument

Like Plantinga, I take naturalism to be the thesis that God does not exist nor anything like God.Footnote ⁴ And, like Plantinga, I take moral realism to be the claim that certain actions and omissions are right or wrong, prohibited or obliged, and that whether or not they are does not depend on human beliefs, desires, emotions, or other evaluative attitudes.Footnote ⁵ Both naturalism and moral realism are theses that have many adherents among philosophers these days. Unsurprisingly, then, there is a substantial number of philosophers who consider themselves to be both naturalists and moral realists. Here, we should think of philosophers like Thomas Nagel, Peter Singer, Erik Wielenberg, and the late Ronald Dworkin. Hence, it would be quite disturbing, if it could be shown that naturalism and moral realism are incompatible. Plantinga aims to defend the thesis that naturalism and moral realism are incompatible, not by providing an argument against the compatibility of naturalism and moral realism, but by rebutting an argument in favour of the compatibility of naturalism and moral realism and by identifying what he considers to be an insurmountable problem for any argument for the compatibility of naturalism and moral realism.Footnote ⁶

Plantinga starts with a preliminary comment. In order to show that two propositions p and q are compatible, one could argue that there is a possible world in which both p and q are true. Such a strategy, however, cannot be employed in the context of naturalism and moral realism. For, if God exists, then, at least on a classical Anselmian understanding of God, he exists necessarily.Footnote ⁷ Hence, if there is some world in which he does not exist, then there is no world in which he exists. To show that there is a world in which naturalism and moral realism are true, then, would also be to show that God does not exist (and, in fact, could not exist). It follows that any convincing argument for the compatibility of naturalism and moral realism would also be a good argument for the non-existence of God and that would surely be too demanding a thing to show.Footnote ⁸

Some philosophers have thought that there is an alternative strategy for showing that naturalism and moral realism are compatible, namely to argue that certain moral properties are equivalent to certain natural properties. An argument along these lines has been defended by Frank Jackson in his From Metaphysics to Ethics.Footnote ⁹ The basic idea is the following. One could ask whether there is some necessarily true proposition of the form ‘An action A is morally obligatory (or morally wrong) iff A has some property P’, where P is a naturalistic property, that is, a property which is such that the fact that it is exemplified does not entail the falsehood of naturalism. It seems that there are indeed such necessarily true propositions. It seems, for instance, that taking some other person's private property without that person's consent is necessarily morally wrong. Moral properties, then, are equivalent to natural properties.

More precisely, it seems that moral properties strongly supervene on natural properties. Properties of kind Φ strongly supervene on properties of kind Ψ if and only if for any possible worlds W and W′ and any objects x and y in W and W′, if x in W coincides on properties of kind Ψ with y in W′, then x in W coincides on properties of kind Φ with y in W′.Footnote ¹⁰ Thus, if two actions A and B have the same natural properties, then it seems they must also have the same moral properties. As Plantinga rightly notes, there are in fact two options here. On the abundant view, natural and moral properties are equivalent (that is, necessarily co-exemplified), but nonetheless distinct. On the sparse view, natural and moral properties are not merely equivalent, but identical: moral properties are natural properties. On both the abundant and sparse views, natural properties are necessarily co-exemplified with moral properties.

But that, Jackson claims, is all the naturalist needs: if moral properties are necessarily co-exemplified with natural properties, then it is plausible to think that naturalism and moral realism are compatible.Footnote ¹¹ Moral properties are or are equivalent to certain natural properties. But there are all sorts of such natural properties, whether or not God exists. Hence, there will be all sorts of moral properties, whether or not God exists. It follows that naturalism is compatible with moral realism.Footnote ¹²

Notice that Jackson's argument does not have the problematic implication that there is some possible world in which both naturalism and moral realism are true, which would imply that God does not exist – for God, if he exists, exists necessarily. It merely implies that certain natural properties are necessarily co-exemplified with certain moral properties. That, as such, does not rule out the existence of God.

Plantinga's response to this argument in favour of the compatibility of naturalism and moral realism is to deny the conditional premise that if moral properties are equivalent to natural properties, then moral realism is compatible with naturalism. In order to see why this premise is implausible, Plantinga asks us to imagine that divine command theory is true. Thus, he asks us to imagine that what makes an action A morally obligatory is that it is an essential property of God to command all persons to perform A.Footnote ¹³ It seems that that is perfectly compatible with the fact that natural properties are equivalent to moral properties. For it could be that natural properties are equivalent to moral properties and that both natural and moral properties are equivalent to supernatural properties, namely if it is an essential property of God to command all persons to perform an action with those natural properties. In such a case, an action with certain natural properties will have certain moral properties in virtue of the fact that it has certain supernatural – and, hence, naturalistically unacceptable – properties. For instance, God could have the essential property that he commands all created persons to treat each other with love and respect. Hence, if one has shown that moral properties are equivalent to natural properties, one has not thereby shown that naturalism and moral realism are compatible.

According to Plantinga, this applies not only when one adopts the abundant view, but also when one adopts the sparse view. On the sparse view, if moral properties are equivalent to natural properties, then moral properties are identical to natural properties. Plantinga asks us again to imagine that divine command theory is true. Then, the action in question will be morally obliged because God has the essential property of commanding created persons to perform that action. Thus, the natural property of being an act of showing love and respect will be identical to the property of being morally obliged because it is also identical to the property of being commanded by God, who has the essential property of commanding it to created persons, if there are any. But surely the property of being commanded by God is not naturalistically acceptable; it would entail the falsehood of naturalism. Hence, even if moral properties are identical to natural properties, it does not follow that naturalism is compatible with moral realism.

Finally, Plantinga points out that the theist faces no such problem: there are good arguments for the thesis that theism and moral realism are compatible with each other. The theist may, for instance, adopt some kind of divine command theory. Thus, she might say that moral properties are necessarily co-exemplified with certain natural properties and that they are necessarily co-exemplified because God commands us to perform actions with certain natural properties, which makes them morally obligated. The theist, then, might maintain that moral properties are equivalent to natural properties, but that this equivalence is asymmetrical: those moral properties depend on those natural properties rather than vice versa, for actions with certain natural properties have certain moral properties because God commands us to perform actions with those natural properties. That certainly seems possible. But if it is possible, then theism and moral realism are compatible with each other.Footnote ¹⁴

A revised argument for the compatibility of naturalism and moral realism

I think there are at least two problems with Plantinga's argument. First, it seems that the naturalist's argument that Plantinga discusses can be revised or at least expanded in such a way that Plantinga's objection is met. It seems that moral properties supervene on natural properties and it also seems that they do so asymmetrically. It is not the case that certain actions have natural properties because we are obliged to perform them, but we are obliged to perform them because they have certain natural properties. As Plantinga rightly suggests, that, as such, does not show that naturalism and moral realism are compatible, because moral properties may asymmetrically supervene on natural properties because the actions which have those natural properties are prescribed by God. Of course, Plantinga does not want to say that that is a metaphysical possibility, that is, that there is a world in which certain actions are obligatory because God commands the created persons in that world to perform that action. For, again, if God exists, he exists necessarily, so that if he exists in one possible world, then he exists in the actual world. Plantinga's response would then entail the existence of God and the falsehood of naturalism, and that is not what he aims to show in his article. What he seems to have in mind is rather that it is epistemically possible that if moral properties asymmetrically supervene on natural properties, they do so because God commands the actions which have those natural and moral properties.

It is not easy to say exactly what it is for something to be epistemically possible. Fortunately, we need not give a full-fledged account of epistemic possibility here. It suffices to note that Plantinga, in his response to the argument in favour of the compatibility of naturalism and moral realism, appeals to it by saying that it is epistemically possible that the scenario in question is one in which moral properties asymmetrically supervene on natural properties because God commands us to perform the action which has those natural and moral properties. Thus, he suggests that for all we know that may be case.

Now, it seems that the naturalist could simply reply by granting that that is epistemically possible, but that an alternative scenario is equally epistemically possible, namely a scenario in which those moral properties supervene on those natural properties not because of a divine command, but simply as a primitive fact, in the same way as it is a primitive fact that the fact that It is true that Friesland has many lakes asymmetrically supervenes on the fact that Friesland has many lakes. For all we know, theism is true and, for all we know, naturalism is true. Similarly, for all we know, divine command theory and moral realism are true, and, for all we know, naturalism and moral realism are true. That would suffice to establish the compatibility of naturalism and moral realism.

But would not the naturalist be unhappy to concede that for all she knows, naturalism and moral realism are incompatible? Yes. But that is not what she concedes. What she concedes is (i) that, for all she knows, moral properties are necessarily co-exemplified with certain natural properties because divine command theory is true, and (ii) that, for all she knows, moral properties are necessarily co-exemplified with certain natural properties because, as a brute matter of fact, the former asymmetrically supervene on the latter. To say that two propositions are compatible – or, at least, to say that we have good reason to think that two propositions are compatible – is to say that there is an epistemically possible scenario in which they are both true. Whether or not there is another epistemically possible scenario in which one or both of them are false is irrelevant to that. This means that we have not only found a good reason to think that, for all we know, naturalism and moral realism are compatible, but that we have also found a good reason to think that they are compatible. For, as I said, two theses x and y are compatible just in case there is an epistemically (rather than metaphysically) possible scenario in which they are both true.

Thus, Frank Jackson's argument can be revised or at least extended in such a way that it meets Plantinga's worry. This means that we have found at least one good reason to think that naturalism is compatible with moral realism.

An alternative argument for the compatibility of naturalism and moral realism

Second, there seems to be another way of showing that two propositions p and q are compatible with each other which Plantinga does not consider. Imagine that there is good reason to think that p does not entail ∼q and that q does not entail ∼p. Take, for instance, the proposition p that Sven is Friesian and the proposition q that Sven likes ice-skating. These two propositions are clearly compatible: the fact that one is Friesian does not rule out one's liking ice-skating – in fact, it seems to render it rather likely. And one's liking ice-skating does not rule out one's being a Friesian – although that, as such, does not render it likely, given the large number of people who love ice-skating and the relatively small number of Friesians on our planet. It seems to me that this is in fact the default way of showing that two propositions are compatible: one shows the compatibility of two theses p and q by showing that there is nothing about the one that entails the falsehood of the other and vice versa.

If we apply this method to naturalism and moral realism, then what is the result? If God does not exist and there is no-one or nothing like God, then does it follow that there are no moral properties that are independent of human values, desires, and beliefs? Or if there are moral properties, then does it follow that God or something like God exists? A problem that arises here is that one might think that these questions are nonsensical, since if God exists, he exists necessarily, and if he does not exist, then his non-existence is necessary. According to the usual way of thinking, all counterfactuals with necessarily false antecedents are true (and all counterfactuals with necessarily true consequents are necessarily true).Footnote ¹⁵ This means that the theist will have to embrace both of the following counterfactuals, for both counterfactuals have, according to the theist, necessarily false antecedents:

• CF₁: If God did not exist, then there would be no mind-independent moral properties.

• CF₂: If God did not exist, then there would still be mind-independent moral properties.

And if both CF₁ and CF₂ are true, then on this alternative attempt to show the compatibility of naturalism and moral realism, naturalism and moral realism are both compatible and incompatible with each other. But that – the theist and naturalist will agree – cannot be true: either naturalism and moral realism are compatible or they are not. Hence, the theist will reject this strategy to show the compatibility of naturalism and moral realism.

It is not clear to me, however, why one should accept the usual way of thinking here, that is, that all counterfactuals with necessarily false antecedents are true.Footnote ¹⁶ Take the following two counterfactuals:

• CF₃: If I were a hippo, I would be able to hold my breath for a really long time and take my naps underwater.

• CF₄: If I were a hippo, I would be able to write philosophy papers.

It seems that CF₃ is true. It seems that if I were a hippo, I would be able to hold my breath for a really long time and take my naps underwater, for that is something hippos can do. However, CF₄ does not seem to be true in this way. Hippos are unable to write philosophy papers. Hence, if I were a hippo, I would not be able to write philosophy papers. This contradicts the usual way of thinking, according to which both CF₃ and CF₄ are true, since both of them have necessarily false antecedents (I could not be a hippo). The usual way of thinking neglects the fact that there is an intuitive difference between CF₃ and CF₄, namely that CF₃ seems true in a way that CF₄ does not, even though both counterfactuals have necessarily false antecedents. If we reject the usual way of thinking about counterfactuals, then we have no reason to think that both CF₁ and CF₂ are true.

If I am right about this,Footnote ¹⁷ then the question whether moral realism could be true if there were no God is not pointless. We can perfectly well ask whether if God does not exist, it follows that there are no objective moral truths. And we can perfectly well ask whether if there are objective moral truths, it follows that God exists. But how should we answer these questions? It seems to me that people's intuitions will widely diverge here. Some naturalists will say that God's non-existence would not rule out the existence or exemplification of moral properties – after all, there would be physical properties (e.g. having a particular length or weight) and, for all we know, all sorts of properties that abstract objects exemplify, so why would not there be moral properties? Other naturalists might feel uncomfortable with the existence of moral properties in the absence of God. Moral properties are weird from a metaphysical point of view: how could it be that certain states of affairs are obliged if nature is all there is? Even theists might have different intuitions here. Divine command theorists might find the idea that there could be moral properties even if God did not exist hard to swallow, whereas intellectualists (those who believe that God deems certain actions right or wrong because they are right or wrong rather than vice versa) might not find that idea so hard to swallow, although they believe that the antecedent is necessarily false.

If this is correct, then the naturalist cannot conclusively demonstrate – or, at least, not convince many fellow-philosophers – that naturalism and moral realism are compatible. But neither is it the case that she must have a defeater for her belief that naturalism and moral realism are compatible, for she might not share the intuition of certain other philosophers that God's non-existence somehow rules out the exemplification of moral properties or the intuition that the exemplification of moral properties somehow implies the existence of God. But if she has no defeater for her belief that naturalism and moral realism are compatible, then, on many theories of rationality (including those of Plantinga himself), the naturalist can rationally embrace both naturalism and moral realism. In fact, the naturalist would have an argument for the compatibility of naturalism and moral realism that would be a good reason for her to think that naturalism and moral realism are compatible. Presumably, it would not convince certain others – intellectualist theists and certain naturalists – but that provides no intellectual problem for that naturalist.

Would not this second line of reasoning in favour of the compatibility of naturalism and moral realism be the very same thing as showing that, for all we know, there is a possible world in which both naturalism and moral realism are true? No. To believe that 3×3=33 and that Affirming the consequent is valid is to believe two necessarily false propositions. If what I said above is right, then there is nothing about the proposition that 3×3=33 that entails that affirming the consequent is invalid, and there is nothing about the proposition that Affirming the consequent is valid that entails the falsehood of 3×3=33. Of course, the propositions that 3×3=33 and that Affirming the consequent is valid are both false. But the problem with believing them is not that they contradict each other or that they are in some other sense incompatible. The problem is rather that there is good reason to think that these propositions are both (necessarily) false. Similarly, the proposition that there are square circles and the proposition that I have drawn a square circle on a paper are both necessarily false, so that there is no possible world in which they are both true, but they seem nonetheless compatible with each other. Thus, to show that naturalism and moral realism are compatible is not to show that there is or might be some possible world in which they are both true, but merely that there is nothing about one or both of these two theses that excludes the truth of the other. The naturalist who thinks that there is nothing about naturalism that excludes the truth of moral realism and that there is nothing about moral realism that excludes the truth of naturalism has good reason to think that the two views are compatible with each other, until someone provides a convincing argument for the incompatibility of naturalism and moral realism.