JR'S Free Thought Pages
The Autonomy of Ethics – David O Brink
David O Brink is Professor of Philosophy, University of California at San Diego.
Preface: In writing this paper, I am conscious of debts to Terence Irwin, Michael Moore, and Sam Rickless. I have also benefited from Michael Moore's "Good without God," in R. George (ed.), Natural Law, Liberalism, and Morality (New York: Oxford University Press, 1996), pp. 221-70. Earlier versions of this material were presented at the Westminster Theological Seminary (Escondido, California) and at a symposium on ethical and religious commitment at the University of San Diego.
Our commitment to the objectivity of ethics is a deep one. Ethics is objective just in case there are facts or truths about what is good or bad and right or wrong that obtain independently of the moral beliefs or attitudes of appraisers. A commitment to objectivity is part of a commitment to the normativity of ethics. Moral judgments express normative claims about what we should do and care about. As such, they presuppose standards of behavior and concern that purport to be correct, that could and should guide conduct and concern, and that we might fail to accept or live up to. Normativity, therefore, presupposes fallibility, and fallibility implies objectivity.1 Of course, this presupposition could be mistaken. There might be no objective moral standards. Our moral thinking and discourse might be systematically mistaken.2 But this would be a revisionary conclusion, to be accepted only as the result of extended and compelling argument that the commitments of ethical objectivity are unsustainable.3 In the meantime, we should treat the objectivity of ethics as a kind of default assumption or working hypothesis.
Many people believe that the only way to make sense of objective moral standards is in terms of divine commands. They assume that moral laws require a lawgiver, such as God, and that a world without God - a purely natural world - would contain no moral standards or distinctions. This assumption explains the frequent appeal to religious scholars and members of the clergy as authorities on issues of moral significance. It also explains why, despite the tradition of the separation of church and state, many people could suppose that constitutional rights depend on divine commandments in the Judeo-Christian tradition.4
This view assumes that morality requires a religious foundation. As such, it denies the autonomy of morality. If an objective ethics presupposes divine command, then an objective ethics stands or falls with religious belief. On the one hand, ethics will be objective if God exists and issues divine commands, and we can acquire moral knowledge insofar as we can know what God has commanded. On the other hand, if theism is false, then the presupposition of an objective ethics fails, and we must embrace moral nihilism (the thesis that there are no facts or truths about ethics) or relativism (the thesis that moral facts and truths are relative to the moral attitudes or beliefs of appraisers). Either way, the assumption that morality requires a religious foundation requires us to reject the possibility of an objective secular morality.
This assumption deserves scrutiny and should be rejected. We should, instead, embrace the autonomy of ethics. The autonomy of ethics implies that the objectivity of ethics is not hostage to the truth of theism. This is a welcome conclusion to the extent that theism is itself a problematic commitment.5 Our discussion assumes not that theism is false, only that it is not obviously true. The autonomy of ethics allows the atheist to recognize objective moral standards. But the autonomy of ethics should not be of interest only to atheists. Whatever its ultimate merits, theism itself is more attractive if we accept the autonomy of ethics. Indeed, a good case can be made that the objectivity of ethics itself requires the autonomy of ethics.
DIFFERENT MORAL ROLES FOR RELIGION
To determine whether morality requires a religious foundation, we need to distinguish three different roles God might play in morality. God plays a metaphysical role in morality if the existence and nature of moral requirements depend on his existence and will. On one such view, it is God's attitudes toward various courses of action that make them good or bad and right or wrong. Second, even if God does not play this metaphysical role, he might play an epistemological role if he provides us with an essential source of evidence about what is morally valuable. Even if God's will does not make something good or bad, it may be a reliable indicator of what is. Third, God plays a motivational role in ethics if he provides us with a necessary incentive or reason to be moral. It is a common view that if we reckon only the earthly costs and benefits of virtue, we cannot show that one is always better off being moral. But if God rewards virtue and punishes vice in an afterlife, then he can provide a prudential motivation for morality.
These three potential moral roles for God all deserve discussion. But our focus should be on whether God plays a metaphysical role in morality. It is this role that has the most direct bearing on the autonomy of ethics.
VOLUNTARISM, NATURALISM, AND THE EUTHYPHRO PROBLEM
We might formulate this metaphysical claim as the doctrine that things are morally good or right just in case God approves of them. In assessing this claim, we would do well to consider Socrates' discussion of a related issue. In Plato's Euthyphro Socrates considers Euthyphro's definition of piety as what (all) the gods love (10a-11b).6 He does not dispute the truth of this claim; instead, he distinguishes two different ways it might be true.
(a) Something is pious, because the gods love it.
(b) Something is loved by the gods, because it is pious.
Socrates does not label these two claims. We might call the first claim voluntarism, because it makes something's piety depend on God's will. We might call the second claim naturalism, because it makes something's piety depend on its nature. Voluntarism claims that the attitudes of the gods make things pious, whereas naturalism claims that something’s piety is part of its nature, which the gods' attitudes recognize and track. Socrates thinks that reflection will show us that we tacitly accept naturalism. What it is for something to be god-beloved is simply for it to be loved by the gods. But what makes the gods take this attitude toward anything must be some other feature of the thing. The gods love pious things, because they are pious. This seems especially plausible if we bear in mind that Socrates and Euthyphro think that piety is part of justice. One loves just things, because they are just, and one recognizes them to be just. If so, the god-beloved character of pious things depends on their being pious, not vice versa. Euthyphro accepts naturalism, and Socrates concludes that Euthyphro's claim fails as a definition, because it states a symptom or correlate, rather than the cause or essence, of piety. Despite superficial differences, Socrates' concern is closely related to ours. We can adapt the Euthyphro Problem to our discussion of whether morality requires a religious foundation by considering the following conditional formulation of the doctrine of divine command.
Divine Command: If God exists, x is good or right if and only if God approves of x.
Divine command, like Euthyphro's definition of piety, admits of both voluntarist and naturalist interpretations, and the debate between them has a distinguished history.
Voluntarism captures the metaphysical dependence of morality on religion. Voluntarists, such as William of Occam (CA. 1287-1347), make something's moral value consist in God's attitudes; there would be no moral attributes but for God's will. Notice that voluntarism and atheism together imply moral nihilism.
By contrast, naturalism accepts the autonomy of ethics. Ethical naturalists, such as Thomas Aquinas (CA. 1225-74) claim that the moral properties of persons and situations depend on their nature. If this is so, moral qualities do not presuppose a God, though a perfectly wise and good God would approve of all and only good and right things. Naturalism does not itself preclude God from playing an epistemic role in morality (telling us reliably what is morally good and bad) or a motivational role (providing divine incentives for moral behavior). But naturalism does deny theism a metaphysical role.
Notice the conditional nature of divine command: If God exists and enjoys the attributes usually ascribed to him - omniscience, omnipotence, and perfect goodness - then he will approve of all and only good and right things. Because naturalism does not make moral qualities depend on God's existence or will, it implies that these qualities would exist even if God does not. Thus, naturalism implies that atheism does not entail nihilism or relativism.
How do we decide between voluntarist and naturalist interpretations of divine command? Socrates' reasoning about piety applies here as well. What it is for something to be God-approved is simply for it to be loved by God. What makes God take this attitude toward anything must be some other feature of the thing. But, it seems, God would love good things because they are good. His attitudes would be principled. If so the God-approved character of good things would depend on their being good, not vice versa.
We might also notice a counterintuitive implication of voluntarism Voluntarism implies that all moral truths are contingent on what God happens to approve. If God's attitudes had been different and he had approved of very different things, then very different things would have been good and bad, right and wrong; and if God were to come to approve of things very different from those he now approves of, then the moral status of these things would change. Thus, for example, had God not condemned genocide and rape, these things would not have been wrong or, if God were to come to approve these things, they would become morally acceptable. But these are awkward commitments, inasmuch as this sort of conduct seems necessarily wrong.7
Theists may reply that God would not approve of such things, because he is himself perfectly good. But this reply is not available to the voluntarist. For this reply understands his approval as resting on a sensitivity to what is good or bad, right or wrong, in itself, independently of his attitudes. But then appeal to God's goodness undermines, rather than supports, voluntarism. Indeed, it looks as if the voluntarist would have to understand God's goodness as consisting in his approving of himself. But that approval would be equally arbitrary and contingent. If, as most theists presumably believe, his self-approval reflects a perception of his own worth, then his attitudes presuppose, rather than constitute, what is of value. The consistent voluntarist account of God's own goodness is problematic.
The way in which the voluntarist must represent moral facts as contingent on God's will is counterintuitive in another way. It is common to believe that the moral properties of actions, persons, institutions, and situations depend in a systematic way on the natural properties - for instance, the biological, psychological, legal, and social properties - of those things. Philosophers refer to this relation of systematic dependence as one of supervenience. The moral properties of a situation supervene on its natural properties just in case a full specification of the natural properties of the situation fix or determine its moral properties. This implies that two situations cannot differ in their moral properties without differing in their natural properties. So, for example, the racial injustice of the system of apartheid supervened on a complex set of legal, political, social, and economic restrictions on the opportunities of black South Africans and a culture of discriminatory attitudes toward them. Any social system qualitatively identical in all natural respects to this system of apartheid would also be unjust, and any social system containing both blacks and whites that was not unjust would have to differ in some of its natural (legal, political, social, economic, and psychological) properties from the system of apartheid. But if the natural properties of a situation determine its moral properties, then its moral properties cannot depend on God's will. For if voluntarism were true, then two situations could have different moral properties even if there were no natural differences between them whatsoever. One system of apartheid could be unjust, but a complete clone of that system need not be unjust - if God's attitudes to the two tokens of the same type were different.8 In this conflict between supervenience and voluntarism, it is easier to accept supervenience than voluntarism.
These are reasons to reject voluntarism and accept naturalism. However, an obvious worry about naturalism is that it compromises God's omnipotence. If moral requirements are independent of and inform God's will, then they are outside his control. But if moral laws are outside God's control, they appear to challenge his omnipotence.
Of course, this is a worry only for traditional monotheists who believe in a personal God who is omnipotent. Atheists and theists of other stripes need not be concerned by this objection to naturalism. Whether traditional monotheists should regard it as a good objection to naturalism depends on how we conceive of omnipotence. If we conceive of omnipotence as the capacity to do anything, then naturalism does compromise God's omnipotence. Theists would then have to choose whether to accept voluntarism (to maintain omnipotence) or sacrifice omnipitence (to avoid voluntarism). But traditional theists face a comparable dilemma anyway. It is difficult to believe that God could change the Iaws of logic (e.g., the principle of non-contradiction) or the truths of arithmetic (e.g., that 2 + 2 = 4). These are necessary truths, true in all possible worlds, and we can't conceive of what a world would be like in which they weren't true. If so, then we already recognize some necessary truths that are beyond God's control. Do they compromise his omnipotence? That depends on how we understand omnipotence. If we understate omnipotence as the power to do anything, then we have independent reason for questioning God's omnipotence. But perhaps omnipotence is not the power to do anything, but rather the power to do anything possible, not inconsistent with necessary truths and laws. If so, God's inability to change laws of logic and mathematics need not compromise his omnipotence. But equally, God's inability to make intrinsic goods bad or intrinsic evils good need not compromise his omnipotence. But then naturalism need not compromise God's omnipotence.
This survey suggests a strong case for rejecting voluntarism and accepting naturalism. Naturalism not only explains how the atheist can recognize moral requirements but also allows theists to explain God’s goodness and to represent his commands as principled, rather than arbitrary. In this way, naturalism appears to be the best bet for atheists, agnostics and theists alike.9 However, unlike voluntarism, which challenges the autonomy of ethics, naturalism vindicates the autonomy of ethics. If we accept the autonomy of ethics, then the objectivity of ethics is not hostage to the truth of theism.
Indeed, we are now in a position to see how the objectivity of ethics actually requires the autonomy of ethics. Ethical objectivity, we said claims that there are moral facts or truths that obtain independently the moral beliefs or attitudes of appraisers. Ethical subjectivism is one way to deny ethical objectivity. It claims that what is good or bad and right or wrong depends on the moral beliefs or attitudes of appraiser. But voluntarism is just subjectivism at the highest level. If God exists and is both omniscient and perfectly good, then his approval - if only we could ascertain it - would be a perfectly reliable - indeed, infallible indicator of what was good or right. This is what naturalism claims. But voluntarism implies that God's attitudes play a metaphysical, not just an epistemic, role in morality; his attitudes make things good or right.
This is a form of subjectivism about ethics. But then the supposition that morality requires a religious foundation, as voluntarism insists, threatens, rather than vindicates, the objectivity of morality.
VARIETIES OF NATURALISM
Naturalism says that moral requirements are not constituted by God's attitudes or will. Put positively, the moral properties of situations depend on the nature of those situations, independently of God's (or anyone else's) attitudes. Indeed, understanding these independent moral properties would help us to understand why God wills what he does (if he exists). But what do moral requirements or qualities consist in if they do not consist in God's attitude or will?
This raises a different inquiry, one within secular moral theory. It is relevant to our inquiry about whether morality requires a religious foundation only insofar as the plausibility of the autonomy of ethics depends on there being some promising accounts of what moral requirements and distinctions do consist in. In developing and assessing such accounts, we necessarily rely on our views about the nature of morality, its demands, and its relation to other concerns.
It might be useful to distinguish moral claims at different levels of abstraction or generality. Some moral claims and judgments are particular. They concern the moral properties of particular actions or action tokens, as in the claim that it would be wrong for Ben to break his promise to Sam. Some moral claims are more general and concern classes or types of action. They identify morally relevant factors and take the form of moral rules, as in the claim that one ought to keep one's promises. Some moral claims are more general still, concerning many or perhaps all types of actions, and saying why these various factors are all morally relevant. These are moral principles. There may be a plurality of moral principles, or, in the limit, there might be only one master principle, such as the principle of utility - which demands that one ought to perform actions that promote human happiness - or Kant's categorical imperative - which demands that one always treat rational agents as ends in themselves and never merely as means.
There is a kind of asymmetrical dependence among the moral claims at these three levels. Particular moral truths (e.g., that it would be wrong to break this particular promise) obtain in virtue of the truth of moral rules (e.g., that promise keeping is a right-making factor). If there are moral truths more general than moral rules - moral principles - then the moral rules express truths in virtue of some more general principle being true that explains why that particular factor is a morally relevant factor.
Some have inferred from this asymmetrical metaphysical dependence that the justification of our moral beliefs must exhibit a parallel structure. They claim that we can justify particular judgments in terms of moral rules and that moral rules can be justified in terms of fundamental moral principles, but that at some point justification must come to stop with either a plurality of principles or a single master principle that states an ultimate moral factor. Because these first principles state ultimate moral factors, there is nothing further in terms of which they can be justified. First principles must be self-evident.
But this leads to an awkward conclusion. We may be uncertain about some particular moral judgments and rules, but surely there are some particular judgments and rules of which we are very certain, much more certain than we are about any recondite first principle. For instance, I am much more certain that the Holocaust was wicked or that genocide is wrong than I am about the truth of utilitarianism or Kant's categorical imperative. Moreover, we need to be able to provide reasons for accepting or rejecting putative first principles.
Fortunately, asymmetrical metaphysical dependence does not imply asymmetrical epistemic dependence. A first principle states a or the ultimate moral factor (e.g., F). It makes no sense to ask of a first principle that we take to be true, "In virtue of what further property is F an ultimate factor?" If there were some more ultimate factor, F would not be an ultimate factor after all. But we can sensibly ask about some putative first principle, "Is that first principle true?" or "Is F really an ultimate factor?" The answer to these questions may appeal to the principle's ability to sustain and explain moral judgments that we find independently plausible. So the metaphysical priority of first principles does not show that our evidence for which first principle is true cannot include our defeasible particular moral convictions.
This conclusion suggests a methodology for secular moral theory. We can try to resolve uncertainty or disagreement at more particular levels of moral thought by trying to find plausible or common ground at a more general level. But we can also try to resolve uncertainty or disagreement at a more general level by testing the implications of a potential moral principle for particular cases against our own independent assessment of those cases. Thus, we introduce a moral principle in order to systematize our considered moral convictions, especially about particular cases and moral rules. We examine candidate principles, in part, by drawing out their implications for real or imagined cases and comparing their implications with our own existing or reflective assessments of those cases. If a principle has counterintuitive implications, this counts against it. But if this counterintuitive implication is fairly isolated, and the principle explains our views better than alternative principles, then this is reason to revise the particular moral judgment or moral rule that conflicted with the principle. Ideally, we modify our principles, considered moral convictions, and other views in response to conflicts, as coherence seems to require, until our ethical views are in dialectical equilibrium.10 On the one hand, dialectical equilibrium begins from and assigns probative value to our considered moral convictions. As such, we have some reasonable expectation that any acceptable theory should accommodate many of our considered moral convictions. On the other hand, dialectical equilibrium is an ideal that none of us now meets and that we can at most approximate. Therefore, we should expect dialectical equilibrium to force some revisions in our moral beliefs, and it is hard to say in advance just how revisionary the moral principles with the best dialectical fit would be.
Secular moral theory should begin with considered moral convictions. For many purposes, appeal to these convictions will be adequate. We have already relied on such convictions - for instance, the prohibitions on genocide and rape - in assessing voluntarism. But if we take the demands of dialectical equilibrium seriously, we must try to identify moral principles that provide a suitable dialectical fit with these convictions. Let me sketch three different theoretical approaches to secular morality.
Much of commonsense morality requires compliance with norms prohibiting aggression (at least, unprovoked aggression), enjoining cooperation, fidelity, and aid, and condemning individuals who free-ride on the compliance of others. We each benefit from the compliance of others with such norms, but others won't be compliant toward those who are known to be noncompliant, and fairness requires that we enjoy the benefits of others' compliance only if we comply ourselves. On this view, we might identify the demands of morality with norms of social behavior the general observance of which is mutually beneficial. This appeal to mutual advantage and reciprocity promises an account of the origin and content of morality that explains our interest in being moral and the interest of the community in instilling a moral sense or conscience in its members.
But this approach appears to limit moral concern to those with whom one regularly interacts. This is in tension with Christian and Enlightenment views that stress the wide scope of moral concern.11 We might understand morality's wide scope as reflecting a perspective that seeks to transcend the agent's personal interests and loyalties. On this view, the moral point of view demands an impartial concern for all. Impartiality can itself be understood in different ways. On one conception, it requires that agents take into account the interests of affected parties equally, balancing benefits to some against harm to others, as necessary, so as to determine the outcome that is best overall. This aggregative conception of impartiality involves a utilitarian or consequentialist approach to morality that identifies one's duty with promoting human happiness or other good consequences.12 However, the aggregative conception lets the interests of many outweigh the interests of a few. An alternative conception of impartiality rejects this sort of interpersonal balancing and insists that a concern with each affected party requires that we act only on principles that no one could reasonably reject. This distributed concern for each yields a contractualist conception of impartiality.13
Still another approach is the Kantian one that attempts to account for the content of moral requirements on the basis of what it is to be a moral agent subject to moral requirements. To be a moral agent is to be responsible. Non responsible actors, such as brutes and small children, act on their strongest desires; or, if they deliberate, they deliberate only about the instrumental means to the satisfaction of their desires. By contrast, a responsible agent must be able to distinguish between the intensity and authority of his or her desires, deliberate about the appropriateness of his or her desires, and regulate his or her actions in accordance with his or her deliberations. If so, moral agency requires capacities for practical reasoning. If moral requirements depend on features of moral agents as such, then they depend on what agents would care about insofar as they are rational agents. This arguably requires a concern for rational agents as such, which is roughly how Kant derives his famous demand that we treat all rational agents as ends in themselves and never merely as means.14
These are mere sketches of a few familiar secular conceptions of moral theory. Our commitment to the autonomy of ethics requires only that some of them look intellectually promising.
MORAL EVIDENCE AND DIVINE WILL
Even if God's will does not make something morally good or bad, it could still be a reliable indicator of what is and provide us with evidence about what our moral duties are. Indeed, if God exists and is morally perfect and omniscient, then his will must be a perfect indicator of what is (independently) valuable. Wouldn't this give religion a significant epistemological role for morality?
Even if God's will provided one source of evidence about morality's demands, it needn't be the only or the most important source. After all, if naturalism is true, then morality's demands have a metaphysical source other than God's will. Moral demands will presumably be a matter of what promotes justice, rights, and happiness. It is open to us to reason directly about these moral matters, by engaging in secular moral reasoning, rather than obliquely by consulting a divine barometer of these matters.
Does divine will provide even one source of moral guidance? Of course, God must exist to provide a moral barometer, and theism is open to serious question. Atheists will think that this direct, secular evidence is all the evidence there is. But even theists should recognize this direct evidence and prefer it if indirect evidence about God's will is sufficiently hard to obtain.
How, after all, are we to ascertain God's will? Appeal to religious tradition and scripture is problematic for several reasons. First, there are multiple traditions and scriptures. Insofar as they say competing things about God's will, they cannot all be true. But it is hard to know how to determine which traditions and scriptures are more reliable. Even within a single religious tradition, questions remain. On some possible moral topics, tradition and scripture may be silent. On other topics, tradition and scripture may speak but in conflicting ways.15 Even when tradition and scripture speak unequivocally, we may wonder whether what is said should be interpreted literally. For example, a literal reading of the Old Testament yields a date for the age of the Earth and claims about the history of plant and animal species that are contradicted by the fossil and geological records.16 It also yields problematic moral claims, such as Deuteronomy's claims that parents can and should stone to death rebellious children (21:18-21) and that the community can and should stone to death any wife whose husband discovers that she was not a virgin when he married her (22:13-21).I7 We have more reason to accept secular scientific and moral claims than we do to accept a literal reading of these particular religious texts.
A common theistic response to these interpretive puzzles is to endorse the interpretation of tradition and scripture that yields the morally more defensible conception of divine will. This moralized approach to interpretation makes good sense for the theist if, as the naturalist claims, God's omniscience and perfect goodness ensure that his will perfectly tracks all morally relevant facts. But, on this conception of interpretation, so far from our knowledge of God's will supplying evidence about the nature of morality, it is our beliefs about the nature of morality that are supplying evidence about God's will. The moralized interpretation of religious scripture and tradition shows religion to be dependent on morality, rather than morality to be dependent on religion.
MORAL MOTIVATION AND THE AUTHORITY OF MORALITY
God plays a motivational role in ethics if he provides a needed incentive to be moral. If we reckon only the earthly costs and benefits of virtue, it appears we cannot always show that one is better off being moral. But if justice requires punishing vice and rewarding virtue, then God's perfect justice seems to imply that he would use heaven and hell to reward virtue and punish vice. Because the afterlife is eternal, its sanctions and rewards would dwarf the earthly costs and benefits of virtue and vice. It follows that the prospect of divine sanctions and rewards could provide a prudential motivation for morality that appears unavailable if we restrict our attention to secular sanctions and rewards.
The question "Why be moral?" is a normative question about why one should care about moral demands. But this normative question cannot be seeking a moral reason to be moral. That question is too easy to answer. Instead, it is asking whether behaving morally is a requirement of practical reason. In this sense, the question really asks about the rational authority of morality. That question arises for most of us because of a perceived tension between the other-regarding demands of morality and a broadly prudential conception of practical reason, according to which what one has reason to do is to promote one's own aims or interests. For meeting the demands of nonaggression, cooperation, fidelity, fair play, and charity often appears to constrain one's pursuit of one's own aims or interests.
When one asks whether virtue pays, one is looking for a prudential defense of the authority of morality. A traditional secular defense of morality is to argue that the demands of morality and enlightened self-interest coincide. As we have seen, much of other-regarding morality involves norms of cooperation (e.g., fidelity and fair play), forbearance, and aid. Each individual has an interest in the fruits of interaction conducted according to these norms. Though it might be desirable to reap the benefits of other people's compliance with norms of forbearance and cooperation without incurring the burdens of one's own, the opportunities to do this are infrequent. Noncompliance is generally detectable, and others won't be forbearing and cooperative toward those who are known to be noncompliant. For this reason, compliance is typically necessary to enjoy the benefits of others' continued compliance. Moreover, because each has an interest in others' cooperation and restraint, communities will tend to reinforce compliant behavior and discourage noncompliant behavior. If so, compliance is often necessary to avoid social sanctions. Whereas noncompliance secures short-term benefits that compliance does not, compliance typically secures greater long-term benefits than noncompliance. In this way, we can provide a secular prudential justification of morality.
However, as long as we understand the prudential justification of morality in terms of instrumental advantage, the secular coincidence between other-regarding morality and enlightened self-interest must remain imperfect. Sometimes noncompliance would go undetected; and even where noncompliance is detected, the benefits of noncompliance sometimes outweigh the costs of being excluded from future cooperative interaction. Moreover, even if the coincidence between morality and self-interest were extensionally adequate, it would be counterfactually fragile. For compliance involves costs, as well as benefits. It must remain a second-best option, behind undetected noncompliance, in which one enjoys the benefits of others' compliance without the costs of one's own. But then, as Glaucon and Adeimantus observe in Plato's Republic (357a-367c, esp. 359c-36id), if one had some way of ensuring that one's own noncompliance would go undetected, one could enjoy the benefits of others' compliance without the burdens of one's own, and one would have no reason to be compliant. The imperfect coincidence of morality and self-interest implies that immorality need not always be irrational.
It is clear that an omniscient, omnipotent, and perfectly good God could arrange eternal sanctions and rewards so as to make the coincidence of morality and self-interest perfect. In this way, appeal to divine sanctions and rewards could strengthen the secular prudential justification of morality. But we might wonder whether morality requires a perfect prudential justification. Perhaps doing the morally correct thing is not always prudent too.
We might also wonder whether this sort of prudential justification of the authority of morality is desirable. For, on this conception of moral motivation, each person has an instrumental justification for being moral, namely, that being moral is both necessary and sufficient for a blissful afterlife. On this conception, moral behavior is good, not in itself, but for its extrinsic consequences. But it is common to think that virtue should be its own reward. Indeed, it is sometimes supposed that when one behaves morally for purely instrumental reasons this diminishes the moral value of one's action.18 God can choose to reward selfless altruism, but the prospect of this reward cannot be what motivates such agents without robbing those actions of the very features he would like to reward.
If virtue should be undertaken for its rewards, this implies a prudential justification of morality's authority. However, if virtue should be undertaken for its own rewards, this justification should eschew appeal to the extrinsic benefits of virtue, which are separable conceptually from the fact of virtue, and appeal instead to benefits that are inseparable from virtue itself. Both the Greek eudaimonist tradition - especially Socrates, Plato, Aristotle, and the Stoics - and the British idealist tradition - especially T. H. Green - defend the intrinsic benefits of virtue, arguing that other-regarding virtues make a constitutive contribution to the agent's own happiness (eudaimonia). Whether these conceptions of the intrinsic rewards of virtue are defensible is a complex matter.19 What is important for present purposes is that if virtue is its own reward, then there will be an important sense in which appeal to divine sanctions and rewards provides a prudential justification of morality that is both unnecessary and unwelcome.
Alternatively, the justification of altruism might eschew prudential mediation, whether extrinsic or intrinsic. This conception of the authority of morality must insist that the fact that I meet some legitimate moral demand of another person is itself a reason for me to act, whether or not I benefit thereby. This would be an impartial conception of practical reason that recognizes non-derivative reason to benefit others. This conception has been most fully developed within the Kantian tradition.20 Whether this conception is defensible is also a complex matter.21 What is significant for present purposes is that this conception of the authority of morality rejects the prudential justification of morality embodied in many religious traditions.
The idea that virtue should be its own reward exposes problems with the adequacy of appeal to divine sanctions and rewards to provide a prudential justification of morality. Whether we decide that virtue is its own reward or that no reward is necessary, it looks as if we can justify moral conduct and concern in ways that assign no role whatsoever to God.
Despite widespread belief that morality requires a religious foundation, that doctrine is hard to support. Voluntarism is the heart of that doctrine and denies the autonomy of ethics. But voluntarism leaves moral requirements unacceptably contingent and arbitrary. It is better to embrace the autonomy of ethics. Indeed, voluntarism is itself best understood as a form of subjectivism. But then it follows that the autonomy of ethics is not only compatible with the objectivity of ethics, but necessary for it. If we accept the autonomy of ethics, then atheism does not force us to moral nihilism or relativism. If we are theists, the autonomy of ethics allows us to explain how God's attitudes and behavior reflect the operation of moral principles. Indeed, it is this moralized conception of the gods that set Socrates apart from the sort of unprincipled polytheism of his ancestors and contemporaries.22 Though the autonomy of ethics denies that moral distinctions rest on God's will, it does not preclude religion from playing other roles in morality. But it is hard to articulate plausible epistemological and motivational roles for God to play in morality. We can see how morality helps religion. It is harder to see how religion helps morality.
1. Different grades of objectivity correspond to different degrees of fallibility. Just how much fallibility and objectivity it is reasonable to expect from ethics is an interesting question, which I won't pursue here.
2. This sort of "error theory" is defended by J. L. Mackie, Ethics: Inventing Right and Wrong (New York: Penguin Books, 1977), chap. 1. Also see Gilbert Harman, The Nature of Molality (New York: Oxford University Press, 1977), esp. chaps. 1 and 2.
3. For systematic defense of ethical objectivity, see David O. Brink, Moral Realism and the Foundations of Ethics (New York: Cambridge University Press, 1989).
4. For example, Chief Justice Roy Moore of the Alabama Supreme Court erected a giant granite monument of the Ten Commandments at his courthouse, claiming that the authority of the Constitution derives from God's commandments. When the monument was ruled to violate the constitutional separation of church and state and was ordered to be removed, Moore refused. He was eventually removed from office for his refusal. See, e.g., "Alabama Panel Ousts Judge over Ten Commandments" New York Times, November 14, 2003, p. Ai 6. More recently, the United States Supreme Court has given mixed signals about whether public displays of the Ten Commandments violate the anti-establishment clause of the First Amendment. In McCreary County v. ACLU, 125 S. Ct. 2722 (2005), in a 5-4 decision the Court ruled that the display of the Ten Commandments by themselves in two county courthouses reflected manifest religious purposes and violated the anti-establishment clause. However, in Van Orden v. Perry, 125 S. Ct. 2854 (2005), in a 5-4 decision the Court ruled that the display of the Ten Commandments as one of many sculptures in a park in the Texas State Capitol did not violate the anti-establishment clause.
5. My own view is that there is no credible evidence for the existence of God, that atheism is the best response to the problem of evil, and that there are perfectly good functional explanations for why people should persist in religious belief despite its falsity.
6. Plato's Euthyphro and other dialogues are collected in Plato: Complete Works, ed. J. Cooper (Indianapolis, Ind.: Hackett, 1997).
7. The issue is sometimes raised in the Judeo-Christian tradition by the Old Testament story in which God asks Abraham to sacrifice his only son Isaac (Genesis 22:1-14). Because Abraham resolves to sacrifice his son, God retracts his command to do so. But this doesn't alter the question of whether God's demanding the sacrifice would have made it right.
8. Voluntarism, like any dispositional theory, which identifies the moral valence of something with its disposition to elicit approval in a suitable appraiser, is committed to a form of moral particularism. I believe, but cannot argue here, that this commitment is further reason for rejecting voluntarism.
9. In "Abraham, Isaac, and Euthyphro: God and the Basis of Morality," in D. Stump (ed.), Hamartia: The Concept of Error in the Western Tradition (New York: Edwin Mellen Press, 1983), Norman Kretzmann endorses many of the doubts expressed here about the viability of voluntarism. But he argues that we can still see God's will as constituting moral requirements if we appeal to doctrines associated with perfect being theology. In particular, Kretzmann appeals to the idea that God's relation to goodness is one of identity, rather than (as usually conceived) predication. I don't fully understand this proposal, but I would note some concerns about it. If God is to be identified with goodness, we might regard this as a reductive identification such that our notion of God is now exhausted by our conception of morality. This would be a revisionary conception of God insofar as it treats God as a moral attribute, rather than a person who might possess moral attributes. Also, it is not clear whether such a view is a rival to naturalism. Someone who was otherwise an atheist and accepts naturalism could accept theism if it amounts to nothing more than recognizing moral goodness.
10. Cf. the method of "reflective equilibrium "in JohnRawls, A Theory of Justice (Cambridge: Harvard University Press, 1971), pp. 19-21, 46-51, and 577-79.
11. The wide scope of moral concern is recognized, for example, in the parable of the Good Samaritan (Luke 10:29-37), who recognizes a duty to rescue someone in need with whom he has no prior association.
12. The best introduction to utilitarianism and consequentialism is still John Stuart Mill, Utilitarianism (originally published 1861), ed. R. Crisp (Oxford: Clarendon, 1998). Some influential contemporary work in this tradition is collected in S. Darwall (ed.), Consequentialism (Oxford: Blackwell, 2003).
13. Important work within the contractualist tradition includes Rawls, A Theory of Justice-, David Gauthier, Morals by Agreement (Oxford: Clarendon, 1986); and T. M. Scanlon, What We Owe to Each Other (Cambridge: Harvard University Press, 1998).
14. See Immanuel Kant, Groundwork for the Metaphysics of Morals (originally published 1785), in Kant's Practical Philosophy, trans. M. Gregor (New York: Cambridge University Press, 1996).
15. Inconsistency is at stake, for example, when we juxtapose the Old Testament doctrine of an "eye for an eye" (Exodus 21:23, 24; Leviticus 24:19, 20; and Deuteronomy 19:21) and the New Testament doctrine of "turning the other cheek" (Matthew 5:38-42; Luke 6:27-31).
16. A literal reading of Genesis implies that Earth was created in six (24-hour) days approximately 6,000-10,000 years ago and that all plants and animals were created at the same time. Geological and fossil records indicate that Earth is approximately 4.6 billion years old, that life first appeared on Earth approximately 3.5 billion years ago, and that many species of plants and animals evolved and, hence, did not exist at the same time.
17. These claims are the tip of a rather large and awkward iceberg of Old Testament passages whose literal interpretation is morally problematic. Cf. (1) Whoever strikes his parents shall be put to death (Exodus 21:15). (2) One cannot kill a slave outright, but one can beat him to death provided that he survives the beating a day or two (Exodus 21:20; but see Exodus 21:26). (3) "Everyone who curses his father or mother shall be put to death... "(Leviticus 20:9). (4) The death penalty is appropriate punishment for various sexual offenses, including adultery and homosexuality (Leviticus 30:10-16). (5) Blasphemy is punishable by death (Leviticus 24:13-16, 23). (6) God instructs the Israelites to stone to death a man found collecting sticks on the Sabbath (Numbers 15:32-36).
18. Cf. Kant's contrast between prudential motivation and a sense of duty in his discussion the good will (Groundwork 390, 393-99).
19. For further discussion, see David O. Brink, "Self-love and Altruism," Social Philosophy and Policy 14 (1997): 122-57.
20. See, esp., Kant, Groundwork, and Thomas Nagel, The Possibility of Altruism (Princeton: Princeton University Press, 1970].
21. For further discussion, see David O. Brink, "Kantian Rationalism: Inescapability, Authority, and Supremacy," in G. Cullity and B. Gaut (eds.), Ethics and Practical Reason (Oxford: Clarendon, 1997).
22. This moralized conception of the gods may also be responsible for Socrates being brought to trial on the charge of impiety, though we may find his moralized conception more pious than its unprincipled rival.