The levers of natural selection are random genetic mutation, fitness for survival, and reproductive success. Defenders of the evolutionary debunking account (EDA) hold that such mechanisms aren’t likely to produce cognitive faculties that reliably form true moral beliefs. So, according to EDA, given that our cognitive faculties are a product of unguided natural selection, we should be in doubt about the reliability of our moral cognition. Let the term ‘sanspsychism’ describe the view that no supramundane consciousness exists. In arguing against theism, some sanspsychists advance a normative claim about the moral significance of phenomena like sentient suffering. But if no supramundane consciousness exists, our cognitive faculties are a product of unguided natural selection. It follows that if EDA is correct, the sanspsychist should not think that our moral cognition is reliable. So unless the sanspsychist has a defeater for EDA, she should not think herself justified in appealing to normative reasons for denying the existence of God.
For what it's worth, I was the commenter on an earlier version of Coley's paper at the 2018 SPR meeting. Below are the comments I presented at that meeting.
Comments on Scott Coley’s “Evolutionary Debunking and Arguments Against Theism”
I want to thank Scott Coley for his stimulating paper on evolutionary debunking and arguments against theism. Coley’s paper argues that Sharon Street’s Darwinian dilemma, when combined with philosophical naturalism, defeats any epistemic grounds for moral realism the naturalist might have had. But if so, then since arguments from evil and divine hiddenness each contain a premise that crucially relies on the truth of moral realism, naturalists cannot consistently appeal to such arguments as grounds for rejecting theism. By bringing important recent work in contemporary meta-ethics to bear on the problems of evil and divine hiddenness, Coley’s argument is a welcome contribution to the dialogue concerning the epistemic credentials of theism and naturalism.
Let us now turn to take a closer look at Coley’s argument.
2. Getting Clear on Coley’s Argument
In laying out Coley’s argument, let’s follow him by letting ‘N’ denote the normative premise in an atheistic argument (such as in the arguments from evil and divine hiddenness); letting ‘C’ denote a conclusion that theism is (at least) probably false; and letting ‘sanspsychism’ refer to any view that entails that no supramundane consciousness exists (e.g., a divine person of some sort). Coley’s core argument then runs as follows:
1. If we affirm sanspsychism then we should embrace moral skepticism.
2. If we should embrace moral skepticism then we should reject N.
3. So, if we affirm sanspychism then we should refrain from affirming N. (1, 2
4. If we assent to C then we should affirm sanspsychism.
5. So, if we assent to C then we should refrain from affirming N. (4,3 HS)
What to make of Coley’s argument? An initial question concerns (1) as currently stated. For Coley grants that at least some versions of moral realism are ultimately compatible with sanspsychism, viz., Peter Railton’s moral naturalism and those relevantly similar to it. Of course we’ve seen that Coley argues that Railton-style moral realism provides no help in supporting N. However, I’m not so sure about that, for reasons I’ll discuss below. Therefore, let us modify Coley’s (1) and (2) accordingly to take account of this other type of meta-ethical option allowed for the sanspsychist. The argument then runs as follows:
1’. If we affirm sanspsychism, then we should embrace either moral skepticism or
Railton-style moral realism.
2’. If we should embrace either moral skepticism or Railton-style moral realism,
then we should refrain from affirming N.
3. So, if we affirm sanspsychism, then we should refrain from affirming N. (1, 2
4. If we assent to C, then we should affirm sanspsychism.
5. So, if we assent to C, then we should refrain from affirming N. (4,3 HS)
The bulk of the paper can then be seen as a defense of premises (1’) and (2’) of the reconstructed argument above.
Given this reconstruction of Coley’s argument, is there any way in which a naturalist might sensibly resist it? In the remainder of the paper, I will sketch some possible replies to the first two premises of the argument on behalf of the sanspsychist.
3. Rejecting (1’)
Start with (1’). Why should we accept it? Here is my reconstruction of Coley’s argument for the premise :
Step 1: If we affirm sanspsychism, then we should embrace moral realism.
Step 2: If we should embrace moral realism, then we should affirm that either selective evolutionary pressures are connected to the moral facts or they aren’t.
Step 3: If we should affirm that evolutionary selective pressures are connected to the moral facts, then we should embrace Railton-style moral realism.
Step 4: If we should affirm that selective evolutionary pressures aren’t connected to the moral facts, then we should affirm that moral realism is defeated by Street’s evolutionary debunking argument.
Step 5: If we should affirm that moral realism is defeated by Street’s evolutionary debunking argument, then we should embrace moral skepticism.
1’. Therefore, if we affirm sanspsychism, then we should embrace either
moral skepticism or Railton-style moral realism.
I can think of four steps in this sub-argument that the non-theist might reject, viz., steps 1, 3, 4, and 5.
Regarding Steps 1 and 5: One might be tempted to object to Step 1 on the grounds that there are plenty of sanspsychists who are living counterexamples to it, e.g., moral relativists, nihilists, and various sorts of non-cognitivists. However, remember that Coley can safely screen these cases off, as (i) he has explicitly restricted the relevant class of sanspsychists to those who want to advance an argument from evil or divine hiddenness, and (ii) (Coley argues that) such sanspsychists can only do so if they also assert both that there are facts about human morality and that human morality overlaps with God’s morality (if God exists). Having said that, however, it’s not clear to me that such restrictions screen off all relevant meta-ethical theories available to the type of sanspsychists at issue. Here I have in mind Kantian constructivist accounts. Versions include those articulated by Kant, Korsgaard, Rawls, and Scanlon. According to at least some of these versions, there are constitutive features of rational, autonomous agents from which a set of categorical imperatives necessarily flows. But if so, then if God is also a rational, autonomous agent, and we are supposed to be made in His image, then perhaps more should be said about why a sanspsychist can’t appeal to such meta-ethical theories to support normative premises in arguments from evil and divine hiddenness.
Regarding Step 3: A number of accounts of moral realism besides Railton’s have been articulated and defended that allow — and sometimes insist — that evolution has reliably shaped our system of evaluative judgments. Perhaps the most relevant is neo-Aristotelian virtue theory, most notably developed and defended by Rosalind Hursthouse, Philippa Foot, and (most recently) Micah Lott. According to this version of virtue theory, moral goodness is analyzed in terms of what is good for members of a given species, which in turn is defined in terms of what allows its members to carry out their characteristic life-cycle, which in turn has been shaped by evolutionary factors. On this sort of account, moral knowledge is a kind of practical knowledge of how to achieve one’s species-specific goods. Given this sort of account, it is guaranteed that evolution will shape our evaluative judgments in a way that is truth-tracking. For those individuals whose evaluative judgments fail to reliably track the truth about what’s good for them are thereby selected out.
Regarding Step 4: A number of non-natural moral realists have rejected this premise. For example, moral intuitionists such as Michael Huemer argue that moral truths are necessary truths, known via reason. Therefore, the way we know moral truths is the same way we know other necessary truths (e.g. those of mathematics and philosophy). And while there is as of yet no complete account of the mechanics of a priori knowledge , this fact hasn’t led many to deny that we have mathematical knowledge. But if not, then by the same token, it shouldn’t lead one to reject moral knowledge.
Is an intuitionist account of moral knowledge metaphysically problematic for the sanspsychist? I don’t see why it would be. That view only rules out supramundane minds, and Platonism doesn’t obviously require that such minds exist.
Perhaps at this point one will raise Benacerraf’s epistemological objection to knowledge of abstract objects: Knowledge requires causal contact with the thing known. But causal contact is impossible with respect to abstract objects; therefore, if moral (or mathematical, or philosophical) facts are abstract, then we can’t have knowledge of them. However, given the certainty and theoretical indispensability of many of these truths, it’s not clear why intuitionists can’t here pull a G.E. Moore shift: “We have reliable belief forming processes regarding mathematical and philosophical truths. If moral facts lack causal powers, then so do these other sorts of facts. Therefore, if moral facts lack causal powers, then our having reliable belief forming processes does not require the facts about which we have beliefs to have causal powers.” (Huemer 2016)
Other (natural and) non-natural moral realists resist this horn of Street’s Darwinian dilemma in other ways. For example, Erik Wielenberg argues that while evolution didn’t select for knowledge of moral truths, it selected for something else that is adaptive and which correlates with moral truths. In particular, he argues that creatures with cognitive capacities like ours thereby have rights. But evolution didn’t select for this knowledge. Rather, it selected for our cognitive capacities, which in turn help us survive and reproduce. But creatures with such capacities are thereby able to grasp the concept of a right, and to come to believe that they have them.
4. Rejecting (2’)
What about premise (2’)? Here is my reconstruction of Coley’s argument for it: Suppose the sanspsychist should embrace either (a) moral skepticism or (b) Railton-style moral realism. Now consider the implications of each option:
Option (a): moral skepticism
Step 1: If the sanspsychist should embrace moral skepticism, then she should (thereby) doubt the reliability of human moral cognition.
Step 2: If she should doubt the reliability of human moral cognition, then she should reject N.
Step 3: Therefore, if the sanspsychist should embrace moral skepticism, then she should reject N.
Option (b): Railton-style moral realism
Step 1: If the sanspsychist should embrace Railton-style moral realism, then she should affirm that she has no good reason to think that our morality is God’s morality.
Step 2: If she should affirm that she has no good reason to think that our morality is God’s morality, then she should reject N.
Step 3: Therefore, if the sanspsychist should embrace Railton-style moral realism, then she should reject N.
2’. Therefore, If the sanspsychist should embrace moral skepticism or Railton-style moral realism, then she should reject N.
What to make of this argument? I can imagine a naturalist raising concerns for the reasoning with both options. First, she might worry about Step 2 for both Option (a) and Option (b). This is because it’s prima facie constitutive of classical theism that God is morally perfect, that God has revealed moral obligations and prohibitions to humans (through special revelation, conscience, and perhaps other ways), and that a number of these apply to God as well. But if so, then to consistently assert the normative premise of the problem of evil, they need merely point out that theism implies that the states of affairs at issue are morally bad, and that God’s moral perfection implies that God can’t permit them without sufficient reason. It therefore seems that even a naturalist who is a moral nihilist could consistently affirm the premises of a version of the problem of evil without committing to the truth of the moral claims involved.
Second, a naturalist might object to Step 1 of Option (b). Here one can reiterate the ways (mentioned in Section 3) in which a naturalist might reason that human morality overlaps with God’s if God should turn out to exist. By way of reminder, they were as follows: (i) If the moral rules are those that any rational, autonomous agent would construct, and God is also a rational, autonomous agent, then prima facie, our morality overlaps with God’s; (ii) the moral rules we’ve discovered apply to at least humans. But if theism is true, then humans are made in God’s image, which presumably includes his likeness in rational and moral respects. There is therefore a presumption for thinking that our morality overlaps with God’s.
I have offered four main replies to Coley’s argument on behalf of the naturalist or sanpsychist: (i) the debunking argument doesn’t rule out Kantian constructivist accounts of meta-ethics; (ii) some forms of moral realism appear to survive the debunking argument (e.g., Railton’s account, Huemer’s account, Wielenberg’s account, and neo-Aristotelian accounts); (iii) there are reasons to think that on any of the accounts above, our morality overlaps with God’s if he should turn out to exist; and (iv) the problems of evil and divine hiddenness can be seen as problems internal to theism itself, in which case the naturalist or sanspsychist need not have substantive moral beliefs to consistently raise them against theism.
i. Assume throughout that ‘morality’ denotes only accounts of morality that lack a divine source.
ii. But see Bonjour (1999), Huemer (2016), Chudnoff (2013), and Bengson (2015) for recent defenses.
iii. Wielenberg (2014), (2016).