The Euthyphro Dilemma, Part 5: Is there a way out?
Recall the two options of the Euthyphro dilemma:
(I) The reason that God commands that we perform morally obligatory actions is that they are morally obligatory.
(II) Morally obligatory actions are morally obligatory in virtue of the fact that God commands that we perform them.
In previous posts in this series I explained what the Euthyphro problem is and why it is a problem. Here is a brief summary of my conclusions: The Euthyphro problem is a problem for option (II), and thereby, a problem for divine command theory. The problem is that if morally obligatory actions are obligatory in virtue of the fact that God commands them, then, since God is omnipotent, his commands are wholly unconstrained. He could command any action whatsoever, and since deontic moral value does not exist prior to his commands, it seems that he has no reason to command one thing rather than another. Indeed, he could command something horrible, such as the gratuitous torture of an infant, and, on option (II), this horrible action would be morally obligatory. But no command can make the torture of an infant morally obligatory. So, option (II) cannot be correct.
As I indicated in the most recent post in this series, there are four distinct aspects of this problem. They are:
(1) The contingency problem
It is possible for God to command anything whatsoever. Given this, no matter what commands God actually issues, it is possible that he issues different commands, even a completely different set of commands (in the sense that, in some possible world, he issues these other commands). So, since on option (II), morally obligatory actions are obligatory in virtue of the fact that God commands that we perform them, there are possible worlds in which different actions are morally obligatory than are obligatory in the actual world; and there are possible worlds in which actions that are morally obligatory in the actual world are not morally obligatory. This is problematic because it seems that at least some moral claims are necessarily true. In particular, it seems that there are some actions such that they have their deontic moral status necessarily. However, if option (II) is correct, then it seems that no action has its moral features necessarily. For any action at all, whatever its actual deontic status, it is possible for it to have a different deontic status.
(2) The counterintuitive possibilities problem
The contingency problem is that, on option (II), all actions have their moral properties contingently. This means that there are non-actual but possible moral truths. A related problem (and a consequence) is that among the possibilities are some that are wildly counterintuitive. The above description of what I’ve called the Euthyphro problem contains an example: it is possible that the gratuitous torture of infants is morally obligatory rather than wrong (in the sense that there is some possible world in which torturing infants is obligatory). But there are other possibilities. It is possible that something morally laudatory, such as giving money to the needy is morally wrong. It is possible that something morally permissible, such as brushing your teeth three times a day, is morally wrong. And it is possible that something morally permissible, such as eating breakfast at 7:30, is morally obligatory.
(3) The arbitrariness problem
If actions are morally obligatory in virtue of the fact that God commands them, then it is difficult to see how God can have reasons for his commands. He has no reason to command one thing rather than another. On option (II), all deontic moral value exists in virtue of divine commands. Deontic moral value is precisely the value that actions have whereby we have reasons to perform or refrain from performing them. Commands are actions. So, if there is no deontic moral value prior to God’s commands, then God can have no reasons for his commands. But if there are no reasons for his commands, then his commands are arbitrary. And if his commands are arbitrary, then morality itself is arbitrary.
(4) The problem of the normative impotence of commands
A command (divine or otherwise) does not seem to be the kind of thing that can make a moral difference. Commands are normatively impotent in the sense that they cannot add to our reasons. We see this when imagine that God commands that we torture an innocent child. This command gives us no reason to torture a child, so it would make no difference to the moral status of child-torture. Or, again, imagine that God commands that we eat breakfast at 7:30 every morning. This command does not give us any reason to eat breakfast at 7:30 am and thus can make no difference to the moral status of refraining from eating breakfast at 7:30 am. Commands are normatively inert in that they cannot add to the reasons that we already have to engage in (or refrain from) the commanded activity. Option (II) thus claims that morally obligatory actions are morally obligatory in virtue of something that is normatively impotent.
Defenders of various versions of DCT have argued that the Euthyphro problem is not a problem for DCT since a properly articulated version of DCT does not have the consequences [(1)-(4)] listed above. In this and the next installment I will offer an assessment of the seriousness of these four aspects of the Euthyphro problem for modern versions of divine command metaethics.
I will consider each of these four aspects of the Euthyphro problem individually. The essay will be broken into two parts. In this current installment, I will look at problems (1) and (2) and the next installment will examine problems (3) and (4).
The contingency problem involves the assumption that it is possible for God to issue commands other than his actual commands (or, stated another way, that it is possible that God issues commands in some possible world(s) that he does not issue in some other possible world(s)). This assumption appears reasonable, at first glance, because God is omnipotent. Given his omnipotence, it seems that God is completely unconstrained; and so it is possible for him to issue any command whatsoever. But, as Edward Wierenga has famously pointed out (along with many others after him), theists do not typically believe that God is completely unconstrained. It is reasonable to believe that God has certain essential characteristics and that, among these characteristics are features that constrain the kinds of motivations that God could experience.
A command is an intentional act of a rational agent. Given this, all commands have motives. If God has certain essential characteristics (or, in other words, an essential nature), this nature will constrain the sorts of motives that God can experience. If that is right (and it certainly seems to be), then it is false that God can issue any command whatsoever.
Wierenga draws our attention to the fact that, on theism, God is perfectly loving. A perfectly loving being, it is reasonable to assume, cannot experience a motive to harm a person who does not deserve to be harmed. If God has this characteristic essentially (as, again, theism implies), then there is no possible world in which God experiences motives to cause harm to a person who does not deserve it. For the very same reason, there is no possible world in which God experiences motives to command that we engage in such horrendous actions as torturing an infant gratuitously.
Thus, as Wierenga argues, theists have a reason to believe that God will not issue cruel commands, namely, the fact that God is perfectly loving. In what follows, I am going to call this argumentative maneuver (that is, the claim that God is constrained by his essentially loving nature), “the appeal to love” (abbrev. ATL). The appeal to love involves the claim that God’s commands are not wholly unconstrained, coupled with the equally important insight that the constraints that apply to God come from within his own nature. Importantly, ATL does not involve claiming that the constraints on God are reason-involving. In other words, the claim here is not that God is constrained to respond to reasons; rather it is that God has a motivational profile and that this motivational profile places limits on the kinds of things that God will do.
While ATL is relevant to both problem (1) and (2), it is very important to distinguish both the problems and the responses. ATL blunts the force of the contingency problem in virtue of the fact that God possesses his loving nature essentially, but I think it is less successful against the problem of counterintuitive consequences.
I say that ATL blunts the force of the contingency problem but, as it happens, the appeal to God’s loving nature is neither necessary nor sufficient to obviate the problem. To show this, I will describe a response that would completely remove the problem:
The contingency problem is completely eliminated if we claim that God has his nature essentially. Given that God’s motives are a manifestation of his essential characteristics, it follows that if he has the same characteristics in all possible worlds, he has the same kinds of motives in all possible worlds. And given the same motives, he will issue the same commands. In other words, the content of God’s character does not matter for the purposes of obviating the contingency problem; all that matters is that God has his nature essentially. Given the way that character grounds motives, if God’s character traits are essential characteristics, then he issues the same commands in all possible worlds.
It is important to see that ATL does not assert this; that is, it does not assert that God has his nature essentially. Rather, it claims that God is essentially loving. But this claim is consistent with the claim that there are other aspects of God’s nature that he possesses only contingently. Perhaps, for example, God is all-loving in all possible worlds, but in some possible worlds he prefers the color red to the color green whereas in others he prefers green to red. Perhaps in some worlds, God enjoys all types of human-produced noise, but in others he has a strong dislike for progressive rock; in some worlds, maybe, he is indifferent to temporal considerations concerning meals, in others he has strong feelings that breakfast must be consumed in the early morning. In asserting only that God is essentially loving, ATL leaves open the possibility that other aspects of God’s character are contingent and thus leaves open the possibility aspects of God’s character may give rise to different motives, and thus different commands, in different possible worlds.
ATL does imply that God’s commands are not wholly unconstrained. Given that God is loving in all possible worlds, it is false that there are possible worlds in which God issues a completely different set of commands in some possible world. But we now see that this does not entirely remove the contingency problem. So long as some aspect(s) of God’s nature is/are unconstrained, then his commands, at least to some extent, will be different in different possible worlds. Given this, DCT, even modified by ATL, will retain the implication that at least some more truths are contingent if it allows that some aspects of God’s nature are contingent.
A brief aside: One might think, at this juncture, that whatever contingency is implied by DCT, it will not be a serious problem. Recall that the most serious problem in this connection is that it appeared that DCT implied that all moral truths are contingent. We have seen that a properly articulated DCT need not have this implication. Perhaps we should not be bothered by the implication that, on DCT, some moral truths are still contingent. It is obvious that not all moral truths are contingent; it is less obvious that no moral truths are contingent. Thus, if DCT implies that some moral truths are contingent, this is not obviously a serious problem for the theory.
I mention this mostly to make sure that I my evaluation of the Euthyphro problem is as thorough as I can make it. I doubt that the contingency issue is the most pressing of the four aspects of the Euthyphro problem and so I will decline to examine the above suggestion any further. My interest, at this point, is to consider what must be added to DCT in order to completely eliminate the contingency problem, and so it is to that issue that I now return.
As I noted above, if God has a nature and has his nature essentially, then, since his motives will be a function of this nature, he will have the same kinds of motives in all possible worlds. Given that commands are motivated actions, this strongly suggests that God issues the same commands in all possible worlds. The assumption that God can issue commands other than his actual commands turns out to be based on a failure to appreciate the ways in which God’s nature informs and constrains his actions. While ATL does not assert this, there is nothing that prevents the defender of DCT from relying on this insight. So long as God has his nature essentially, it seems that DCT does not imply that morality is contingent.
Before moving on to problem (2), I want to make a couple of observations that will be very important to the evaluation of problems (2) – (4): First, the fact that God has an essential nature implies only that he experiences certain kinds of motives rather than others; it does not imply anything about what those motives are. If God’s nature is cruel, then he will experience cruel motives and issue cruel commands; if he is loving, then he will experience compassionate motives and issue compassionate commands.
Second, and perhaps more importantly, while the response that I have sketched to the contingency problem involves the claim that there are constraints on God’s commands, these claims about God’s nature (including ATL) do not assert that the constraints on God are reason-involving. In other words, the claim here is not that God is constrained to respond to reasons; rather it is that God has a motivational profile and that this motivational profile places limits on the kinds of things that God will do.
The contingency problem is entirely eliminated if we assert
(EDN) Essential Divine Nature: God has his nature essentially.
But this will not resolve the counterintuitive possibilities problem since it is consistent with EDN that God’s nature is cruel or indifferent. A god with a cruel nature will command that we torture innocent children. We might think that ATL entirely resolves the problem, but that would be an error. To see why, note that the problem of counterintuitive possibilities is not merely the problem that God might command something cruel. In the above description of the problem, I mentioned four different kinds of counterintuitive possibilities:
(A) God might command that we perform some cruel act(s), such as the torture of an infant.
(B) God might command that we refrain from performing some compassionate act(s), such as helping to feed people who are starving.
(C) God might command that we perform some seemingly neutral act(s), such as eating breakfast at 7:30 am.
(D) God might command that we refrain from performing some seemingly neutral act(s), such as eating dinner at 6:30 pm.
The appeal to love can only address possibilities of type (A) and (B). A perfectly loving God would neither command a cruel act nor command that we refrain from compassionate acts. However, ATL cannot address possibilities of type (C) and (D). That God is loving does not imply that he will not command that we eat breakfast at 7:30 am. At least, it is very unclear how such a command would be inconsistent with his love. It is wildly counterintuitive that it could be morally obligatory to eat breakfast at 7:30 am. But DCT seems to imply that it is possible that doing so is morally obligatory and ATL provides no reason to think that this is not a genuine implication.
How serious is this problem? Since I am not claiming here that, on DCT, it is possible that it is morally obligatory to doing horrible things, it may seem that the problem is not all that serious. But I think it would be a mistake to dismiss the problem posed by possibilities of type (C) and (D). Every day, hundreds of times a day, you and I engage in actions that are seemingly innocuous but are such that it would not be inconsistent for a loving God to command that we not do these things. We drink coffee without cream and sugar, we brush out teeth after breakfast but before leaving the house for the day, we read news articles online while we are eating breakfast, we watch sporting events on television, we tell jokes, we converse with friends and co-workers at work, etc., etc. It would not be inconsistent for a loving God to command that we refrain from performing any of these things. If God commanded that we not read news articles while we eat breakfast, then according to DCT, it would be morally wrong to read a news article during breakfast. But it is wildly implausible that it could be morally wrong to read a new article while one eats breakfast. The same can be said for all of the other items on my list and countless other seemingly innocuous actions that people engage in every day.
In a subsequent post, I will return to the issue of seemingly innocuous actions in connection to problem (4) (the problem of the normative impotence of divine commands) where I think it has a great deal more force. For now let me close this aspect of the inquiry by pointing out that the appeal to love does not eliminate all of the counterintuitive possibilities that appear to be consequences of DCT.
 This is only partially accurate. Usually such defenders focus on (1) – (3) to the neglect of (4). In addition, many such defenders confuse the distinct aspects of the problem (e.g., it is common to confuse the contingency problem with the arbitrariness problem). I will attempt to substantiate these claims in a future post.
 Wierenga, E. “A defensible divine command theory.” Nous 17.3 (1983): 387-407.
 Other philosophers and defenders of DCT have offered responses to the Euthyphro problem that are similar to Wierenga’s. I will be discussing some of these related responses in future installments of this series. It is perhaps worth pointing out that while this current essay does not involve a thorough discussion of each and every argument that defends DCT from Euthyphro-type concerns, my assessment of the seriousness of the Euthyphro problem for DCT is an overall assessment. That is, parts 5 and 6 of this series contain my all-things-considered assessment of the Euthyphro problem.