Recall Adams’ answer to the famous arbitrariness objection to DCT: God’s command is not arbitrary in the contemporary pejorative sense because it commands what is good, but it has discretion over which good things to require. To make this picture work we need an account of goodness. Chapter 1 suggested that a person who says a thing is good expresses that she is drawn or attracted by it and says that it deserves to draw or attract her in that way. But now we need an account of the criteria for this deserving. Adams proposes a single criterion for when “good” means “excellent”: a thing is good to the extent that it resembles God. But while this fits many kinds of excellent things, it does not fit all. In particular it does not fit the goodness of natural kinds. Take two different species of plant that are different; there is resemblance to God in life, but also in such differences, but the differences aren’t differences in likeness to God. There’s a different goodness in the way a particular plant within a species differs from another. This is what follows from Scotus’s picture that we get more perfection as we move from genus to species, and then yet more as we move from species to each individual member, to the haecceity. Resemblance to God does not seem well equipped to explain all this difference in the good.

Here are four plausible ways of thinking of the relation of this difference to God. First, God created it. This suggests not simply that God is the source, but that God delights in the variety; perhaps this is one of the reasons for the “very good” in Genesis 1. Second, these various beauties draw us to God. Chapter 1 claimed that what finally draws us and deserves to draw us is God and what draws us to God. This makes the account dynamic rather than static, and that is an improvement, because it connects better with the account of the meaning of “good,” which is based on the idea of being drawn. But the goodness of the variety does not reside only in its drawing us to God. It also does not reside only in its having been created by God, which would make the value too transparent to God and insufficiently intrinsic. The third way of thinking is metaphysically the most ambitious. Perhaps we learn about God from the creation in the same sort of way we learn about an artist from his or her work. It’s not that the features resemble him, but the work is, nonetheless, utterly characteristic. It manifests his aesthetic preferences and aesthetic personality, which is in a sense present in the work. Two different plant species might manifest God in slightly different ways. This is close to Adams, but different. Manifestation and resemblance are not the same because of this sense of God’s presence in the manifestation. A fourth picture might be that each kind of flower species, indeed each plant, is a part of the biotic whole, which God loves. This is different from the previous picture because it makes the value of the parts derive from the value of the whole. We can add that God intends a whole new heaven and earth, not just new humans as co-lovers, but a new creation. Perhaps non-human species have goodness in their destination just as individual humans do, though there are large difficulties with this view, such as dinosaurs.

The moral law is good for humans because it fits human nature. Scotus says the precepts in the second table of the Ten Commandments are “exceedingly in harmony” with the first practical principles that are “known from their terms,” even though they do not follow from them necessarily. Scotus is saying that the second table fits our end, but is not deducible from it. He standardly describes this relation of “fit” in aesthetic terms. This account of goodness allows us to give some content to the way in which the good, at least the good for humans as humans, overridably constrains what we should take to be obligatory.

This suggestion is partly the same and partly different from the one Geach made in his famous article about the good. He distinguished “predicative” adjectives like “red” from “attributive” adjectives like “good,” on the grounds that the meaning of attributive adjectives can’t be detached from the meaning of the nouns to which they are applied. He held that “good” in phrases like “good human” is a purely descriptive term, and we know its meaning just by knowing the meaning of “human” and “human act” and so by knowing what humans and human acts are for.

Hare concedes that we do get a constraint on what we should take to be good for humans from knowing what humans are, and that this in turn constrains what we should take to be commanded by God. But Hare wants to make two points against Geach. First, the term “good” in “good human” and “good human act” has, in our ordinary language, more than a purely attributive use. For a Kantian a good human act is one that displays a good will, and this is defined in terms of the whole procedure of the Categorical Imperative. Hare thinks we should concede that “good” in “good person” is sometimes used attributively, and that in “good human being” it may be used both ways. Second, even the attributive use is more than merely descriptive. To ascribe a function to a chisel is normally to say how it is to be used, and to say that a chisel is to be used for cutting fine grooves in wood and is not to be used as a screwdriver is to make a prescription.

Adams’ DCT needs a constraint on what we should take to be commanded by God in order to overcome the objection from arbitrariness. If God’s command makes something obligatory, should we think that we might be commanded, and thus obliged, to fly a plane into a skyscraper? No, says Adams, for this could not be the command of a loving God. What counts as “loving” is settled by ordinary valuation, for Adams, but Hare thinks this seems wrong, for reasons he’ll give later. But a related idea seems right: we should probably not take something as commanded by God if it does not fit the characteristic kind of loving of God done by a rational animal. This would imply a presumption against taking an innocent human life, though it wouldn’t rule out such a command and at the same time God changing the circumstances. Also, recall various ways Abraham’s context is different from ours (no life of Christ to emulate, for example). At any rate, taking innocent life is inconsistent with the feature of our rational-animal agency that features each of us going through a particular trajectory.

This point comes helpfully into connection with the four Barthian constraints mentioned earlier. Our trajectories involve our being individual centers of agency, being in time, being free, and being language-users. This is often put in terms of our personal narratives. We don’t know in advance the good works God’s prepared for us, and God is free with respect to the route selected for each of us, and the duration of that route. God has discretion, but chooses in accordance with our good as pilgrims. But the question of what access we have to the nature of this good is a different one, though also an unavoidable one. It’s natural to assume the trajectory requires a full lifetime, so killing would be a premature termination, unless God were to indicate to the contrary.

This gives us one example of a constraint from our nature on what we count as good, if we are to be able to respond to the divine call to be co-lovers. There is a form of argument here that can be extended to other examples. The form of argument is transcendental in the Kantian sense; it argues from the conditions of possibility of some fact that is taken as basic. Thus the argument from providence is a transcendental argument from the “fact of reason,” that we humans (creatures of sense and creatures of need) are under the moral law. The present argument argues backwards from the fact that we are recipients of God’s call to conclusions about what we (and God) have to be like to make this possible. But then it reverses direction, and asks what constraints are placed on what we can take to be divine commands by the fact that we are this kind of people.

Now consider the proscription on bearing false witness. There is a plausible argument, this time from the last of the four Barthian constraints, that we are language-users. Our language is a system of external signs, used to communicate internal thoughts. But then we have to be able to assume that these signs are being used most of the time to communicate thoughts correctly. What is at issue here is not whether the thoughts communicated are true, but whether what is communicated corresponds to what is thought true. The plausible part of Kant’s argument is that the institution of language-using requires that we be able to trust that most of our fellow language-users are communicating what they believe to be the truth, or what they really want. So there is a presumption against telling a lie, because if anyone, anytime, with any degree of frequency, may lie, this undercuts the very institution of language, which is necessary for lying as well as for telling the truth.

In the case of lying, as in the case of killing the innocent, we are left with an overridable constraint. Consider by contrast the exceptionless kind of constraint offered by the “new” natural law theory, which proposes eight basic goods, as Porter puts it, “elemental enough to be regarded plausibly as self-evident to all and yet provided with enough content to provide an immediate basis for practical reflection.” The dilemma that Porter poses for this kind of theory is that either the list is sufficiently general to be self-evident, but then it does not have enough content significantly to guide action in the exceptionless ways the theory proposes, or it is specific enough to guide action in this way, but then it is not self-evident to all. To pose this dilemma, however, is not to deny that goods such as life and truth pose some constraint on moral obligation.