John Hare’s God’s Command, Chapter 4, Section 4.1.3, “The Social Character of Obligation”:

Hare now asks if we, by bringing in human nature in this way, have abandoned the distinctive mark of divine command theory, and simply turned it into a species of natural law theory. The argument has not been that the moral law is natural law strictly speaking, but that the content of at least two of the Ten Commandments has been turned into presumptions against taking God to be commanding us to act in a certain way, and these presumptions are taken from what fits human nature. For Scotus, the first table is natural law strictly speaking (except for the “seventh day” prescription). The command to love the neighbor would also be natural law strictly speaking, since we are necessarily commanded to love God, and to love the love of God, and therefore to love the neighbor’s love of God. But Scotus believes in the possibility of reprobation, so there is a restriction needed: we are commanded to love the love of God in the neighbor “at least by anyone whose friendship [God] is pleased to have.” We can and should have a defeasible presumption that we and the neighbor are not among the reprobate, because the judgment is God’s and not ours. Moreover, since we are necessarily commanded to love God, and since human nature is specified in terms of this end, we can say that God necessarily commands what fits human nature. But Scotus does not think that any of the specific commands in the second table can be deduced from this. There are two different possible kinds of deduction from what fits us. There is a deduction of a presumption in two cases, but in no case is there a deduction of an absolute prohibition.

A second point is more important. The argument so far doesn’t imply the moral law or moral obligation is deducible from human nature even in the case of the prohibitions on killing the innocent or on lying. There can be a presumption against doing something and still not an obligation not to do it. Here we return to Adams and Darwall and their notion of the social character of obligation, which we can accept with one qualification. The social character is that we are obligated to someone, or by someone. [Murphy objects that, whereas tort law always has a tortfeasor and a victim, and so has a “bipolar” structure, this is not true of criminal law, which can have a “monadic” structure in which there may be no victims at all (God and Moral Law, 126). If this is right, we should not say that obligation as such is bipolar. But there is good reason, Hare thinks, explored by Darwall in The Second-Person Standpoint, ch. 5, to think that moral obligation is more like tort law in this respect.] The opposite of “obligatory” is “forbidden.” It is not at all an easy matter to delineate this social character, but the general point seems right. The qualification is that we should not derive the agent’s obligation from the goodness to the agent of the relation that would be damaged by violating the obligation. That would be another form of eudaemonism. But that aside, suppose we start with the way Adams puts the basic idea, that, where there is a violation of an obligation, one “may appropriately have an adverse reaction to it.” The question is: Who is it whose appropriate reaction is here in question? Human beings have limited information, and limited sympathies. Even if we did know the preferences of others, we would tend to prefer the preferences of some people to the preferences of others in a way not countenanced by the moral law.

We might ask, why should we assume that the person to whom we are accountable in an obligation is the same as the person who generated the obligation in the first place? There is a tradition of argument, in Kant, for example, and also in Suarez, that God is legislator, executive ruler, and judge, and that moral law assumes that it is the same person who carries out all three functions. This tradition lay behind the discussion of God’s authority in Ch. 2. In Kant’s terms, the author of the law (which we repeat in our own wills) has to have a holy will, the administration of the law has to be by the “supersensible author nature,” and the judge has to be able to see into our hearts; and there is one person, and it is the same person, who does these three things. We might ask: “Why could it not be three different persons?” After all, in human societies it can be an advantage to have these functions divided. Hare is content to make this modest claim: If there is only one God, that one God is the most appropriate person for these three roles.

If this argument works, or something like it works, we can say that moral obligation requires not just a presumption against doing something but an obligator, and that deducibility from human nature and non-divine facts alone therefore has to be denied. Now we can return at last to Murphy’s dilemma. The first horn proposed that, if God is free to command what God wants, the non-moral and non-divine facts are inert. But if, as Hare has argued, God is constrained though not determined by facts about our nature, these facts will not be inert. The second horn of the dilemma proposed that it is odd to say that we are obligated not by the maximal set of non-moral, non-divine facts (where these include facts about our nature), but God is so constrained. The response is that this is not odd at all. We and God are different. Both God and we are constrained by non-moral and non-divine facts, and neither God nor we are obligated by those facts. But we are obligated by God’s commands. God does not require an obligator at all, but is the obligator. Even in those cases of moral law (if there are any) in which God’s command is constrained by the non-moral, non-divine facts, we are obligated not by those facts but by God’s command.