The Euthyphro problem

The arbitrariness objection is widely attributed to Plato's Euthyphro. In the Euthyphro, Socrates asks Euthyphro to define piety/justice (hosion).Footnote ² Euthyphro proposes the following definition: ‘What is dear to the gods is pious, what is not is impious’ (Plato (2000), 7a). Socrates rejects this definition because he claims that it leads to a contradiction. According to the Ancient Greek religion, the gods often disagree – what some gods love other gods hate. Therefore, some things are both loved by some gods and hated by other gods. It follows that some things are both pious and impious, but that is impossible. Euthyphro concedes this objection (Plato (2000), 7e–8a). At Socrates' suggestion, Euthyphro endorses the following revised definition: ‘the pious is what all the gods love, and the opposite, what all the gods hate, is impious’ (Plato (2000), 9d–9e). This revised definition avoids Socrates' objection to the first definition since it doesn't commit one to endorsing contradictions (saying that certain acts are both pious and impious) in cases in which the gods disagree. In cases in which a particular act is loved by some gods and hated by other gods, the second definition implies that the action in question is neither pious nor impious. (However, the second definition raises another problem that is not mentioned in the dialogue and, to my knowledge, has not been noted by anyone else – if the gods disagree about everything or almost everything, then Euthyphro's definition implies that nothing, or almost nothing, is pious or impious. In this connection, it is important to ask how many gods ‘all the gods’ refers to – if the list of the relevant gods is very long and includes minor deities, then the likelihood of agreement among all the gods is greatly diminished.)

Socrates begins his criticism of this second definition by asking, ‘Is the pious being loved by the gods because it is pious, or is it pious because it is being loved by the gods?’ (Plato (2000), 10a). He asks Euthyphro to choose between the following:

1. (A) The gods love what is pious because it is pious.

2. (B) What is pious is pious because the gods love it.

Clearly A and B are inconsistent and Socrates and Euthyphro both assume this. More problematically, they both assume that one of these alternatives must be true.Footnote ³ Euthyphro answers Socrates' question by endorsing A and rejecting B. He doesn't explain or defend this answer nor does Socrates ask him to. Socrates seems to agree with Euthyphro's answer. Socrates then argues that Euthyphro's answer to the question is inconsistent with his definition of piety.

Socrates' argument against Euthyphro's second definition of piety can be summarized as follows:

1. 1. If Euthyphro's (second) definition is true/correct, then B is true (and A is false).

2. 2. B is false (and A is true).

Therefore,

This argument is valid (it has the form of a modus tollens argument). Note that the argument is valid if premise 2 reads simply ‘B is false’.Footnote ⁴ Therefore, if both premises are true, the conclusion must be true. Premise 1 is clearly true; it is true by definition. If being pious means being loved by all the gods (or if piety is the property of being loved by all the gods), then what is pious is pious because it is loved by all the gods. The only serious question about the argument is whether premise 2 is true. Socrates doesn't give a serious argument for thinking that A is true (and B is false). He asks Euthyphro to choose between A and B. Euthyphro answers that A is true, but he doesn't give any reasons for this. Socrates agrees with this answer so he doesn't ask Euthyphro to defend it. But clearly this assumption needs to be defended. This glaring weakness of the Euthyphro argument is masked by the fact that the Euthyphro is written as a dialogue.

This problem notwithstanding, many contemporary philosophers side with Socrates; they think that B is clearly false. Here's the argument. The gods must have some reason for loving and hating the things they love and hate. Otherwise, their loves and hates are arbitrary. If the gods' loves and hates are arbitrary, then there is no reason to take them seriously as the ultimate standard for morality. This argument assumes that if B is true, then the loves and hates of the gods must be arbitrary. But this assumption is false. Given Euthyphro's definition, we can't say that the gods love what they love because it is pious. This rules out one possible way in which the loves and hates of the gods could be non-arbitrary. But the conclusion of this argument, that if B is true then the loves and hates of the gods are arbitrary, follows only if we accept something like the following:

This statement presents a false dichotomy. Given his definition of hosion, Euthyphro can't say that the gods conform their loves and hates to some independently existing standard of hosion. But this leaves open many other possible reasons why the gods might love some things and hate others. There are all sorts of different reasons one can have for loving or hating something. So, at best, the Euthyphro argument is incomplete, and, if we extend it in the way analogous to many recent arguments to the effect that divine will/divine command theories make God's will arbitrary (see Footnote endnote 1), the argument clearly fails. However, even if we grant all of this, it still might be the case that the kinds of reasons that the gods could have for loving and hating things (given the truth of Euthyphro's theory) don't give us reasons to regard their loves and hates as a moral standard that is binding on us. The key question is this: given the kinds of reasons that the gods could have for loving and hating the things they do that are compatible with the truth of Euthyphro's theory, do we have reasons to conform our will to the loves and hates of the gods? (I will have much more to say about this point later.)

With only slight modifications, Socrates' arguments apply to monotheistic analogues of the positions discussed in the Euthyphro. Consider the following monotheistic analogue of Euthyphro's second definition (I will call this ‘the divine attitude theory of right and wrong’):

Socrates' crucial question can be modified as follows:

This question asks us to choose between:

1. A′. God loves right acts because they are right.

2. B′. Right acts are right because God loves them.

The modified version of Socrates' argument can be stated as follows:

1. 1′. If the divine attitude theory of right and wrong is true, then B′ is true and A′ is false.

2. 2′. A′ is true and B′ is false.

This argument is valid (the conclusion follows from the premises), and premise 1′ is clearly true. (The argument is also valid if we modify 2′ to read simply ‘B′ is false’.) The only thing in doubt is premise 2′. Why think that premise 2′ is true – why think that A′ is true and B′ is false?

If we follow the argument for A over B presented earlier, the argument would be roughly this: either we must say that God loves right acts because they are right or else God has no reason whatever for loving and hating the things he does (or at least no reason that can appropriately ground moral/value obligations). This argument has the same problems as the earlier argument. It rests on a false dichotomy – it is possible that God loves or hates actions for reasons other than their being right or wrong.

A modified version of Socrates' argument also applies to the DCT. The DCT holds the following:

The analogue of Socrates' question for this theory is:

This question asks us to choose between:

1. A″. God commands/permits/forbids us to do what is obligatory/permissible/wrong, because it is obligatory/permissible/wrong.

2. B″. What is obligatory/permissible/wrong is obligatory/permissible/wrong because God commands/permits/forbids it.

The DCT commits us to accepting B′ and rejecting A′. If we endorse A′ (or if we say that B′ is false), then we must reject the DCT. The argument for accepting A″ and rejecting B″ is similar to the argument for accepting A over B (and accepting A′ over B′). The argument is roughly as follows: either we must say that God commands/forbids right/wrong actions because they are right/wrong or else God has no reason whatever for commanding/forbidding what he commands/forbids. This argument is open to the same objections noted earlier.

The Euthyphro argument has a glaring weakness; it gives us no reason to accept Euthyphro's answer to Socrates' question. I have formulated a subsidiary argument for Euthyphro's answer analogous to recent arguments against the DCT, but that argument is also seriously flawed. Surely it is possible that the gods have reasons to love/hate what they love/hate, even if there is no standard of piety independent of the gods' loves and hates. Similarly, it is possible that God has reasons to command/will what God commands/wills, even if there is no independent standard of right and wrong. But this alone doesn't save divine will/divine command moral theories from related objections. It might still be the case that the kinds of reasons God could have for willing/commanding what God wills/commands that are consistent with the truth of divine will/divine command moral theories provide us with no reason to defer to God's will or obey God's commands. The key question is what we have reason to do in light of God's willing/commanding that we do certain things. More precisely, given that God's reasons for willing/commanding what God wills/commands don't include God's being moved by independent moral facts about what is right and wrong, can the fact that God wills or commands us to do certain things give us reasons for thinking that God's will/commands is/are binding on us? A closely related but distinct question is this: how can God's reasons for commanding/willing certain things (which according to divine will moral theories can't be moral reasons that involve God's recognition of independent moral facts) create moral reasons for us?

Different versions of divine will/divine command theories give very different accounts of what sorts of reasons God can have for willing/commanding as God does and, therefore, we need to look at the details of particular divine will moral theories and ask if they can provide us with good reasons to follow God's will/commands. We cannot answer these questions in the abstract without considering the details of particular versions of divine command/divine will moral theories. I now turn to five recent divine will/divine command moral theories and their answers to the Euthyphro objection.

Adams's three modified divine command theories

In his paper ‘A modified divine command theory of ethical wrongness’, Robert Adams defends a DCT of right and wrong, but holds that there are standards of good and bad, personal welfare, and virtue and vice that are independent of God's commands (Adams (Reference Adams1973), 114–116). Adams writes:

Adams can't say that God commands what he commands because it is morally right (independently of being commanded by God). But Adams is not committed to the view that God's commands are arbitrary or that God has no reason to command one thing rather than another. Adams can say that God commands what God commands because God is loving, omniscient, and desires our good; because God loves human beings, God desires what is good for us.Footnote ⁶

Adams presents his second modified DCT in ‘Divine command metaethics modified again’. He writes:

According to Adams, Atheists who reject the DCT understand the concept of wrongness, but, unbeknownst to them, the best theory about the nature of wrongness is that wrongness is the property of being contrary to the commands of a loving God. Similarly, our pre-scientific ancestors all understood the concept of water but they didn't know the correct theory about the nature of water – that water is H₂O (Adams (Reference Adams1979), 136).

In ‘Divine command metaethics modified again’, Adams says nothing about standards of good and bad and seems to continue to assume that there are objective standards of good and bad and welfare that are independent of God's will and commands. At least everything he says there is consistent with his holding that there are independent standards of good and bad. Thus, although he doesn't address the arbitrariness objection in this paper, his second theory is consistent with exactly the same reply to the arbitrariness objection that the first version of his theory permits. He can't say that God commands what he commands because it is morally right (independently of being commanded by God). But he is not committed to the view that God's commands are arbitrary or that God has no reason to command one thing rather than another. Adams can say that God commands what he commands because God is loving and, because he is loving, God desires what is good for us.

Can Adams give a satisfactory answer to the second question I noted near the end of the section on the Euthyphro problem? If there are no independent moral facts about what is right and wrong that help constitute God's reasons for commanding what God commands, how can God's commands create moral reasons for us? Adams has a perfectly good answer to this question. He can say that the fact that a loving God commands us to do X for certain reasons (together with the fact that God's reasons include independent facts about what is good and bad and independent facts about what is good and bad for human beings) creates moral reasons for us to do x. In this way, God's commands (together with those independent facts about what is good and bad and what is good and bad for human beings) can create moral facts about right and wrong.Footnote ⁷

In his more recent book Finite and Infinite Goods, Adams proposes yet a third version of his modified DCT. He states his theory of right and wrong as follows:

He qualifies his theory by saying that he endorses it only ‘on the assumption that’:

Adams stresses that he endorses the DCT as ‘a theory of the nature of moral obligation only, and not of moral properties in general’ (Adams (Reference Adams1999), 250).

The theory of right and wrong Adams defends in Finite and Infinite Goods is roughly the same as the theory he proposes in ‘Divine command metaethics modified again’. However, Finite and Infinite Goods presents a substantially different theory of value than Adams's earlier writings. In this book, Adams defends a Platonistic theory of value according to which God is perfectly good and other things are good to the extent that they resemble God (Adams (Reference Adams1999), 14).

The two earlier versions of Adams's theory can answer the arbitrariness objection by appealing to the idea that God's commands are motivated and shaped by God's love and his concern to promote the good, where what is good is determined by standards of value independent of God. Since the divine resemblance theory of value is a component of the most recent version of his theory, Adams can no longer say that God's reasons for making certain commands rather than others include God's being moved by standards of value independent of God. This raises the question of whether Adams's most recent formulation of his theory can answer the arbitrariness objection.

Adams holds that, in order to count as the form or standard of the Good, God must have certain features; a deity who was cruel or who loved cowardice could not count as the form of the good (Adams (Reference Adams1999), 46).

Adams says that God's commands cannot be the source of moral obligations unless God is just. Adams's DCT relies on what he calls a ‘thin theory of justice’ which ‘does not presuppose moral obligation as such’.Footnote ⁸ According to Adams, standards of justice are independent of God's commands and God's will, but not independent of God's nature. God is the constitutive standard of goodness/excellence and justice is one of God's excellences. Things are good/excellent to the extent that they resemble God, but Adams does not say that every possible omniscient, omnipotent deity would constitute the standard of goodness:

This leaves us with questions about the criteria for determining whether a particular deity is or is not a suitable candidate for being the ultimate standard of goodness. Surely these criteria (which in Adams's theory are more basic than theories of good and bad and right and wrong) must be independent of God's actual nature. In Adams's theory, the meaning of conventional moral concepts is the starting point and foundation for everything else. He makes a sharp distinction between our conventional moral concepts and metaphysical theories about the nature of moral properties. He says that his divine resemblance theory of good and bad and his DCT of the right and wrong are the theories that best account for the role assigned to goodness and rightness by our conventional moral concepts (Adams (Reference Adams1999), 15–18 and 250–258).

As I argue below in my discussion of Zagzebski, the concept of love presupposes the concept of welfare – a necessary condition of loving someone is that one desires the welfare/good of the other person. This raises the question of whether Adams's theory involves some kind of vicious circularity. Adams claims that something's being good or good for someone consists in its resembling a loving God and that the property of being morally obligatory is identical to the property of being commanded by a loving God. But the statement that God is loving presupposes a prior independent notion of welfare/goodness; see below, where I argue that this kind of vicious circularity is a serious problem for Zagzebski's theory. To the extent that this is a serious problem for Zagzebski, it is very likely a serious problem for Adams as well. But I won't insist on this point. I am much more concerned to show that I have a good answer to the arbitrariness objection than to show that the third version of Adams's theory does not. If Adams's most recent theory is viciously circular, he can avoid this problem by modifying along the lines of my theory – see below.

In the two earlier versions of his theory, Adams defends a hybrid theory according to which some, but not all, moral notions are determined by God's commands. These two theories appeal (or can appeal) to realist standards of goods and bad, welfare, and virtue and vice independent of God's will. Clearly divine will/divine command theories of morality that anchor standards of good and bad as well as standards of right and wrong in God's will/commands cannot make use of the quick and ready answer to the Euthyphro objection that is open to defenders of the DCT who think that standards of good and bad are independent of God, and, to that extent, such theories are much more vulnerable to the arbitrariness objection.

Zagzebski's divine motivation theory

Zagzebski defends what she calls ‘exemplarism’. She defines moral concepts such as right and wrong and good and bad in terms of the motivations of moral exemplars or paradigmatically good people. Moral theorizing, she contends, should begin by pointing to and identifying moral exemplars. Identifying exemplars is prior to having a theory about what makes someone an exemplar. Similarly, identifying paradigm cases of gold or water is prior to analysing or discovering the essence of gold or water. She states two versions of her theory, a naturalistic version that doesn't presuppose the existence of God and a full theory (her preferred version of the theory) for Christian philosophers that presupposes the divinity of Christ. Her full theory claims that God (Jesus) is the perfect moral exemplar. She defines fundamental moral concepts in terms of the exemplar's (God's) motivations.

To have a good life is to have the kind of life that the exemplar would want for herself (Zagzebski (Reference Zagzebski2004), 116 and 134). An act is right (permissible) provided that: (1) it is not the sort of act that an exemplar would characteristically refrain from doing in like circumstances, and (2) an exemplar would not feel guilty for doing it in like circumstances. An act is wrong provided that an exemplar: (1) would characteristically not do it in like circumstances and (2) would feel guilty about it if s/he did it. An act is a moral duty provided that: (1) an exemplar would characteristically do that act in like circumstances, (2) an exemplar would insist on doing it, and (3) an exemplar would feel guilty for not doing it.Footnote ⁹

Zagzebski claims that her theory gives a good answer to the arbitrariness objection, but that Adams's theory and my divine preference theory do not.

According to Zagzebski, arbitrariness is a problem for traditional versions of the DCT, which imply that we have a duty to follow God's commands even if he commands cruel and hateful acts (Zagzebski (Reference Zagzebski2004), 259–260). Adams's theory doesn't have this consequence, since if God were to make such commands he would show himself to be unloving and thus, according to Adams, we wouldn't be obligated to follow God's commands in that case. She contends that Adams's answer to the Euthyphro objection is still unsatisfactory because it is ‘ad hoc’.

In this respect, Zagzebski's theory may be preferable to the DCT on grounds of theoretical simplicity, but this is hardly a decisive objection to Adams's theory. It is unclear why it is a serious problem for Adams that he ties morality to two distinct properties of God instead of one. (Adams appeals to the commands of a loving God; Zagzebski appeals to God's loving motives). More needs to be said here. However, I won't dwell on this point, because if my defence of my own theory against the arbitrariness objection works, it can also be extended to defend Adams's theory against the objection – see the discussion of my divine preference theory below.

Zagzebski claims that her own theory completely avoids the problem of arbitrariness:

If her theory avoids the arbitrariness objection, that would be no small thing. It would show that divine will moral theories can answer the Euthyphro objection even if they don't hold that standards of good and bad are independent of God and God's will/commands.

Does Zagzebski's answer to the arbitrariness objection succeed? I think not. Zagzebski to the contrary, not every motive is a justifying reason. There are bad motives. The motive of hatred (or unjustified hatred) is not a justifying reason. Here, Zagzebski could reply that the motive of love is a justifying reason and that everything God does is motivated by his love. Even if everything God does is motivated by love,Footnote ¹⁰ it is not so clear that love is always a good motive, i.e. a motive that provides justifying reasons. Not every act motivated by love is a good action that is worthy of imitation. Wrong acts and stupid acts can be motivated by love. Murders and other crimes are sometimes motivated by love. Why couldn't God's loving motives go awry in similar ways? Although Zagzebski doesn't address this question, a promising answer is available to her. In cases in which wrong actions are motivated by love, e.g. the parent who harms another person's child in order to help her own child win a competition, the agent fails to show love for those she harms. Inasmuch as God loves everyone (equally), his loving motives can't be objectionable in the way that human actions motivated by very partial and limited love can be.

Even if we accept her claim that God's love is (always?) a justifying reason, there is another serious problem for Zagzebski. Her answer to the arbitrariness objection depends on her appeal to the idea that God is loving and motivated by love. According to her, what is good and good for people is determined by the motives of a loving God. This creates serious problems for Zagzebski because it commits her to rejecting standard definitions of love, according to which a necessary condition of loving a person is that one desires that person's good/welfare. Standard definitions of love make the notion of good (or someone's good/welfare) conceptually prior to the notion of love.Footnote ¹¹ If the standard definitions of love are correct, it is viciously circular to say that what is good (or good for someone) is determined by the motives of a loving God. This is a problem that I raised earlier for my own divine preference theory of value (Carson (Reference Carson2000), 246–248). Zagzebski is aware of this problem and quotes from my statement of the problem (Zagzebski (Reference Zagzebski2004), 221). The passage she quotes from me is as follows:

Zagzebski attempts to defend an alternative account of love that avoids this problem. She writes:

Zagzebski claims that one can have love for something that has no welfare. One can love art, philosophy, or chocolate. But this is strained. We don't love persons in the same way we love inanimate objects or activities. These examples involve different senses of the word ‘love’. Zagzebski claims that love consists in having a certain kind of emotion towards an object. Love is an emotion that is expressed in certain characteristic ways. Like pity and anger, love has characteristic ends (Zagzebski (Reference Zagzebski2004), 222). God does not aim at our good qua good. Because he loves us, God wants us to be healthy and have friendships, etc. (Zagzebski (Reference Zagzebski2004), 222)). God also wants us to be happy, not suffer, and be full persons (Zagzebski (Reference Zagzebski2004), 320). God desires those things because that is what loving beings desire for those they love (Zagzebski (Reference Zagzebski2004), 222).

According to Zagzebski, desiring the good of another person is an effect of love, not a defining characteristic of loving that person. Love is an emotion that leads to such things as a desire for a person's welfare, but concern for the other person's welfare is not necessary for love. This is doubtful. Suppose that I feel what Zagzebski calls the emotion of love for someone but my emotion doesn't have its usual causal consequences, and I fail to have any concern for her welfare (or any concern for the things that constitute her welfare). Shouldn't we conclude that I don't really love the other person? I won't pursue this further here. As Zagzebski observes, everything depends on the proper definition of love – ‘the real issue is what love consists in’ (Zagzebski (Reference Zagzebski2004), 221). However, since she rejects the standard view, Zagzebski has the burden of proof – a burden she hasn't fully met.

My divine preference theory

In Value and the Good Life, I defend a divine preference theory of value by appealing to a divine preference theory of rationality. I argue that what is good is what it is rational to desire and that (on the assumption that the ‘right’ sort of God exists) the best theory of rationality is that it is rational (in the sense of ‘correct’) for someone to desire what God prefers that she desire. I didn't propose a theory of right and wrong action in Value and the Good Life. In fact, I believe that if a suitable God exists, then right and wrong are in some way determined by God's will, but I haven't defended a preferred version of the divine will theory of right and wrong.Footnote ¹²

The preferences of a cruel, hateful deity are not a plausible standard of good and bad. In order for God's will to constitute the standard of good and bad, God must have certain virtues or psychic traits. I hold that if a suitable God exists, then God's preferences are the standards of good and bad, but, if such a God doesn't exist, then some other standards determine what is good and bad.Footnote ¹³ Here is a summary statement of my view from Value and the Good Life:

According to my theory, if a kindly and omniscient (etc.) God exists, there is no standard of good and bad independent of God's will. Since I think that love needs to be defined in terms of desiring the other person's good, I can't say that what is good depends on the will of a loving God. That would be viciously circular. I need to describe the kind of God whose preferences determine what is good/bad without any reference to God's being good or loving. I describe God as an omniscient person who created the universe and human beings for certain purposes/reasons, and is caring, kind, sympathetic, and unselfish. If Zagzebski is right about the concept of love (and I very much hope that she is), then I can modify my theory to say that it is the preferences of a loving omniscient God that determine the goodness and badness of things. However, if she is not right about this, then her theory has the same problem as mine. In that case, she can't appeal to the idea that God's motives are loving, and she may have to modify her description of God's motivations along the lines of my theory. (These comments also apply to Adams's theory. If his theory is also viciously circular in this way, he can avoid this problem by modifying it along the lines of my theory.)

Zagzebski poses the following dilemma for me:

Zagzebski says that if God's preferences are determined by his underlying motives, e.g. God's love/kindness, then my theory is really a divine motivation theory, and she doesn't object to it. But she insists that God's motivations provide a deeper explanation of why things have the moral properties they have. God's motivations provide ‘the rock bottom features of the divine psyche out of which arise obligation or moral correctness’ (Zagzebski (Reference Zagzebski2004), 269). Zagzebski is correct that motives often determine preferences and commands. She is also correct to claim that, when God's motives determine his commands and preferences, his motives constitute a deeper level of explanation of moral concepts than his commands or preferences. However, I strongly disagree with what she says about cases in which God's preferences are not completely determined by his loving motives.

I accept the second horn of the dilemma she poses (I think that there probably are cases in which God's preferences aren't fully determined by his underlying loving/kindly motives), but I don't accept her view that this saddles me with a serious objection about the arbitrariness of God's will. It seems possible that God could have preferences that aren't determined by his underlying virtues (his kindness and sympathy, etc.). Consider a case in which God's virtues and essential nature underdetermine his preferences for our choices and preferences. Suppose that God's preferring that I prefer A to B and God's preferring that I prefer B to A are both equally compatible with and equally expressive of God's underlying kindness, sympathy, and knowledge, etc. It is possible that there are many such cases. Many different sorts of preferences seem to be compatible with the description of God that I use in my theory. On the other hand, it is also possible that God's purposes might hang together in ways we don't understand and that many things that seem arbitrary to us are not.

Suppose that God prefers that I prefer A to B (and that I choose A over B), but this preference doesn't derive from any of God's virtues. Let us grant that this implies that God doesn't have any reason to prefer that I prefer A to B (or choose A over B). Nonetheless, I still have reason to prefer A to B (and choose A over B). Even if God's preferences are in some sense arbitrary, they give us reasons to conform our preferences to God's preferences concerning them. Divine preferences for which God has no reasons can generate reasons for us. Sometimes friends and loved ones have whims that involve wanting us to do or prefer certain things. For example, suppose that, on a whim, my father wants me to prefer X to Y (or choose X over Y). Suppose that independently of this, it is not objectionable for me to prefer X to Y (or choose X over Y). This gives me some kind of reason to have the preferences that he wants me to have. My reason(s) might include reasons of gratitude. I love the Chicago Cubs and strongly desire that they play in the World Series. My having this desire is optional. Even though I do not have any good or compelling reasons for having this desire, my having it gives my wife some sort of reason to prefer or desire that the Cubs play in the World Series. My wanting her to want or do something gives her some reason to want or do it. It is not a problem for my view that it implies that moral facts can be created by arbitrary divine preferences (preferences God has for which he has no reason).Footnote ¹⁶

But there is a serious problem with the forgoing. On my view, something is intrinsically good provided that it is rational for people to have an intrinsic desire that it occur or be the case. But my examples don't show anything about the rationality of intrinsic desires/preferences. They only show that the arbitrary desires of others can give us reasons to have instrumental desires for things they want us to desire. In the case of my wife and my desire that she desire that the Cubs win the World Series, my desire gives her reasons to desire that the Cubs win the World Series as a means to my happiness and as a means to pleasing me, but it does not give her any reason to have an intrinsic preference that the Cubs win. If God has an arbitrary preference that I prefer X to Y, that gives me a reason to prefer X to Y as a means to pleasing God, but it does not give me reason to have an intrinsic preference for X over Y.

Here is my answer to this objection. What it is correct/rational for me to desire is what God wants me to desire. What it is correct for me to desire as an end is what God wants me to desire as an end. There may be things that God desires/prefers me to desire but that God does not desire me to desire as ends. God might simply want me to desire these things as a means to pleasing God and deferring to God's judgement. In such cases, God's preferences don't give me reasons to have intrinsic preferences for the things that God wants me to desire. But if God desires that I desire X for its own sake, then I have reason to desire X for its own sake, and X is intrinsically good. If God desires that I intrinsically prefer X to Y, then it is correct for me to prefer X to Y for its own sake, and X is intrinsically better than Y.

In order to adequately address Zagzebski's thought experiments about cases in which God has desires that aren't fully motivated by his virtues (his kindness/lovingness, etc.), it would be helpful to contrast my theory with what I call the ‘ideal observer version of the divine preference theory of value’ (IODPT). The IODPT can be stated as follows:

God's being kind and omniscient underdetermines God's preferences. These characteristics are compatible with different kinds of psychological characteristics and inclinations. There is no reason to think that all possible kindly (loving?) omniscient deities would have the same preferences about everything.Footnote ¹⁷ In light of this, we probably should revise the IODPT by saying that X is good/bad provided that all possible kindly omniscient Gods would prefer that X exist/not exist.

Several standard versions of the ideal observer theory (IOT), including those defended by Roderick Firth and Richard Brandt, hold that the reactions of an ideal (but human) observer determine the truth or falsity of normative judgements. These theories hold roughly the following: a favourable (unfavourable) moral judgement about X is true provided that, if s/he existed, an ideal human observer (or all human ideal observers) would have a favourable (unfavourable) attitude about X. These versions of the IOT do not posit the actual existence of an ideal observer. IODPT is similar to the ideal observer theory in that it takes the preferences that God would have (if God existed) to be the ultimate standard of rationality. There may be truths about what God would prefer (if God existed), even if God doesn't actually exist. According to the IODPT, what things are good or bad is independent of whether God actually exists. The IODPT differs from Firth's and Brandt's versions of the IOT in that it takes God, rather than a super-duper human being, to be the ideal observer. By contrast, my DPT makes God's actual preferences the ultimate standard of rationality, only if a loving and omniscient creator of the universe actually exists.

I think that we should prefer my DPT to the IODPT. The following scenario makes clear what is at stake between my theory and the IODPT. Suppose that there is a kindly (loving?) omniscient creator of the universe who has clear preferences about how human beings live their lives. Suppose also that certain other possible kindly (loving?) omniscient creators of the universe would have different preferences about how humans should live their lives. According to my theory, God's actual preferences determine what it is correct to prefer (and thus what is good and bad). But the IODPT does not imply that God's actual preferences have any priority over the preferences of other possible Gods. In such a case, we should accept the implications of my DPT and say that what is correct for us to prefer is determined by God's actual preferences. Human beings have reasons to harmonize their preferences with God's. But aren't God's actual preferences arbitrary since they could have been different without God's losing any of God's perfections? God might have had different preferences from those God actually has. God might have designed the cosmos and our place in it for different purposes from those God has in fact designed them for. God might not have created the earth or human beings at all – God might have chosen to create other forms of intelligent life instead. But the contingency of (at least some of) God's actual preferences and designs does not alter the fact that we have reasons to submit ourselves to those plans and preferences. Given the existence of actual divine purposes within a particular plan of creation (which has its source in a kindly omniscient being), we have reasons to fulfil those purposes. If there exists a kindly omniscient (etc.) creator of the universe who has a plan for the universe in which human beings play a role, then it makes sense for us to fulfil our purpose in that plan. If such a God exists, then it is reasonable for us to try to live in accordance with God's plan and God's preferences and not resist our natural ends and the design of the universe. It would be reasonable for us to defer to the superior knowledge and wisdom of such a God, just as it is reasonable for small children to defer to the wishes of their parents, provided that the parents are loving. Our own vantage point for judging normative moral questions is seriously deficient.Footnote ¹⁸

In this connection, I need to raise an earlier question in connection with my view. Since, on my view, there are no moral facts of any sort that are independent of God's will, none of God's reasons for preferring or commanding that I do one thing rather than another can be moral reasons. How can God's non-moral reasons create moral reasons for us? Very briefly, the answer is that God doesn't have independent moral reasons to will as he wills, but God's willing as God wills creates moral reasons for us to do as God wills. The fact that a kindly, omniscient creator of the universe wants/commands me to do X gives me an overriding reason to do X. God's preferring/commanding that I do X creates/constitutes the moral fact that my doing X would be good/right. Granted that God's preferring/commanding that I do X creates a reason (perhaps a strong reason) for me to do X, why think that those reasons are overriding? The answer is that there is no one else (including me) whose standpoint for judging these things is nearly as good as God's. God has perfect knowledge and a full understanding of how what I do fits with the purposes for which God created the universe. To the extent that my own preferences might conflict with the preferences God wants me to have, it is reasonable for me to defer to God's will and God's superior standpoint for judging these things. Because of this, and, very importantly, because there are no independent moral facts that could give us reasons to defy God's will,Footnote ¹⁹ it is reasonable for us to defer to God's will to the extent that we can discern it.