There have been two general approaches to the connection between God and morality in the history of Western philosophy. According to what is usually called voluntarism, God created morality by an act of his free will and imposed it upon us. For us, morality is obedience to a divine sovereign; for God, morality does not clearly apply at all. According to intellectualism, there are eternal moral truths that God did not create and that are valid for God as well as for us. God commands us, but God's commandments come from his intellect's knowledge of these truths. In both approaches, morality is law. In Part II, I will propose a third alternative. God is essential to morality, not because it comes from either his intellect or his will, but because it comes from his motives. God's motive dispositions, like ours, are components of his virtues, and all moral value derives from God's motives. There are many exemplars of goodness in ordinary life, but the ultimate paradigm of goodness and the source of value is God. The moral properties defined in Part I ultimately refer to the motives of God.

THE INTELLECTUAL PROBLEM OF EVIL

The most serious challenge to the belief that there is a perfectly good and omnipotent God is the problem of evil. This problem confronts theists both intellectually and existentially. The intellectual problem is that the existence of evil seems to make the existence of a God with these attributes improbable or even impossible. It is therefore an attack on the rationality of belief in God. The existential problem is that seeing and enduring evil makes it hard for humans to trust God even if they come to the conclusion that His existence is not unlikely and that theistic belief is not irrational. The existential problem attacks religious believers' motivation to go on with their lives in a way that gives them meaning. Philosophers are primarily concerned with the intellectual problem, and I will concentrate on that problem in this chapter, but the existential problem is so important that it should never be very far in the background, and although appeal to philosophy is far from the best way to deal with the latter problem, some philosophical positions help more than others.

STARTING WITH EXEMPLARS

If a natural kind N is defined by direct reference, it follows that the metaphysical question “What is N?” is not independent of the epistemological question “Can we identify an N?” Direct reference ensures that our semantic community cannot be radically mistaken about the members of the class of N. Water just is the class of stuff like this. Since the semantics guarantees that there is a sense in which we know what N is, it protects against a certain kind of skepticism. We have already seen that this does not mean that we can identify what makes something an N. That is something we have to find out by scientific investigation. But even about that question, we have a safeguard against a form of skepticism. Whatever makes water what it is, it is what makes this kind of stuff what it is, so at least we know what we need to investigate in order to find out the nature of water.

Exemplarist ethical theory protects against a form of moral skepticism for the same reason that direct reference protects against a form of skepticism about natural kinds.

Divine Motivation theory gives virtue ethics a foundation in a theological metaphysics of value. It has a number of advantages over the generic form of motivation-based virtue ethics outlined in Part I, and I would like to conclude this part of the book by comparing them. In motivation-based virtue ethics, emotions have intrinsic value or disvalue in a way that parallels the intrinsic value or disvalue of beliefs. Beliefs are good when they fit what they are about, and when they do, we call them “true.” Emotions are good when they fit what they are about, although we do not have a word for a good emotion other than “fitting” or “appropriate.” There is a direct connection between good emotions and true beliefs, because when an emotion is good, the standard proposition expressing it is true. The intentional object of the emotion has the thick property that the agent sees it as having in the state of emotion.

In addition to beliefs and emotions, there are other things with intrinsic value. I have briefly discussed pleasure and mentioned that epistemic states other than true belief probably have intrinsic value – for example, understanding. I have not ruled out the possibility that many other states of persons have intrinsic value, and there may be objects other than persons that are good intrinsically, although I have not discussed them in this book.

Intrinsicality pertains to the source of the value of a thing, not to its degree.

THE PROBLEM OF MORAL DISAGREEMENT

One of the most difficult problems for moral philosophy is unresolvable moral disagreement, either between persons in the same culture or between persons in different cultures. This problem is modern in origin. Clearly, that is not because disagreement has only recently been discovered, but because it was not previously thought to be a problem. The difference in attitude reflects an evolving view on the nature of the self, a change from which there is no turning back. We now commonly view the beliefs of others as being prima facie as rational as our own, so if our considered judgment is that they are not rational, we need a reason to think so. The mere fact that our beliefs are ours and theirs are theirs is not a sufficient reason to think that ours are more rational than theirs. And any reason to which we can point in support of a judgment that our beliefs are more rational than theirs all too often can be turned against us, sooner or later. The result is a sense of the fragility of our beliefs, including the moral beliefs that ground our acts, and some of these beliefs are closely connected to our sense of self. We are more conscious of self than in previous eras, largely because we are more conscious of the threats to it.

Introduction

I will argue that rational satisficing and ethical satisficing suffer from the same problem, namely, that they are in danger of making no sense because their conceptions of the good cannot be made out independently of a conception of the proper aim of actions. In the case of rational satisficing, the main argument comes from the foundations of decision theory. In the case of ethical satisficing, I borrow an argument from Philippa Foot.

In the end it turns out that ethical satisficing survives the problem, whereas rational satisficing does not. I will motivate the distinction by appealing to an intuitive feature of commonsense morality, namely, supererogation.

Rational Satisficing

The literature on rational satisficing is loaded with interesting examples, but none of them has ever struck me as an example of rational satisficing. They always appear to be examples of something else, not always of the same thing, but always of something that defenders of maximizing could accommodate within their theory. One of these is the example of Hannah. Suppose Hannah has put her house on the market, and naturally she wants the best price for it. Because she can't wait indefinitely for bids to come in, she decides to accept the first bid that comes in above a certain satisfactory price that she chooses in advance. Suppose she picks $200,000 as her threshold. If someone offers her $205,000 the first day the ad appears in the local real estate listings, she'll take it.

It is testimony to the breadth of thought of Herbert Simon, the man who conceived the idea of ‘satisficing’, that the concept has influenced such a wide variety of disciplines. To name a few: Computer science, game theory, economics, political science, evolutionary biology, and philosophy have all been enriched by reflection on the contrast between choosing what is satisfactory and choosing what is best. Indeed, these disciplines have cross-fertilized one another through the concept. So one finds satisficing computer models of evolutionary development, satisficing economic models of international relations, satisficing applications of game theory within economics, and philosophical accounts of all of these.

Philosophical interest in the concept of satisficing itself represents a convergence. The fecund and appealing idea of choosing what is satisfactory finds a place in the theory of practical reason, or thinking about what to do. The appeal of the concept derives partly from the fact that what is satisfactory is, well, satisfying. Satisfaction is generally good, and goods of this generality feature prominently in any account of practical reason. More noteworthy is the fact that the concept of satisficing finds application from so many perspectives, even within the relatively narrow confines of moral theory.

In any conversation of this complexity it is always in point to ask whether the participants are talking about the same thing. So of course this issue arises with respect to the essays collected in this volume.

Introduction

I understand the problem of satisficing to be that of whether it is morally permitted (at least sometimes) for agents to choose an action that is less than the best when better actions are in their power and known by them to be so. The primary motivation for accepting the moral permissibility of satisficing is to reduce the demandingness of ethics. The “demandingness objection” has traditionally been directed at consequentialism, but I shall relate the discussion to the demandingness of virtue ethics.

The demandingness objection to certain moral theories is the objection that these theories contain requirements that are too demanding on agents. There are a number of possible strategies for overcoming the objection. A vulnerable theory may allow that it is morally permitted to satisfice, or given that it is not morally permitted to satisfice, it may propose a criterion of best action that is not itself too demanding. A weaker reading of the demandingness objection suggests a third strategy: It is possible that a theory may possess a demanding criterion of rightness or requirement in that sense, but a less demanding conception of conditions under which agents can be blamed for failing to perform right acts, even when those acts are in their power and known by them to be so. Let me briefly elaborate on those three strategies.

The first strategy replaces maximizing or optimizing criteria of rightness with satisficing criteria.

Satisficing theories, whether of rationality or morality, do not require agents to maximize the good. They demand only that agents bring about outcomes that are, in one or both of two senses, “good enough.” In the first sense, an outcome is good enough if it is above some absolute threshold of goodness; this yields a view that I will call absolute-level satisficing. In the second sense, an outcome is good enough if it is reasonably close to the best outcome the agent could bring about; this leads to what I will call comparative satisficing. These two views coincide in their implications for a specific sort of case, in which the situation is now fairly far below the absolute-level threshold and an agent can at best bring it to a point somewhat above that threshold. Here both absolute-level and comparative satisficing say that one need not bring about the best available outcome, though of course one may; one is required only to improve the situation to the absolute threshold. But in other cases the views diverge. If the situation is now far below the absolute threshold and, no matter what, will remain below it, absolute-level satisficing requires agents to do everything they can to improve the situation; here its implications coincide with those of maximizing. But comparative satisficing is less demanding, requiring agents only to make some reasonable percentage of the largest improvement they can.

When we think about whether it ever makes sense to choose something that is simply good enough even when other better things might instead be chosen, it seems that it can. We can readily be offered examples in which it seems that is just the choice we sensibly make. To use one oft-cited example, it makes sense to accept a reasonably good offer on a house one is selling rather than hold out for a higher price. It may be just as reasonable to hold out for more, but provided the offer is good enough, there is nothing irrational or unreasonable in accepting the first sufficiently good offer.

Thus the examples seem to show that satisficing — that is, to choose the merely good enough over an option which is better yet — is sometimes rational. However, many philosophers have wanted to argue that things are not as they seem. They wish to defend the idea that satisficing is rational only if it serves as part of an overall strategy to maximize. In service of this position, they have available a general strategy for dealing with examples that purport to show otherwise. This strategy is to ask the advocate of satisficing what it is about the lesser option that justifies one's choosing it over the greater. Once a reason is offered, the clever proponent of maximizing can incorporate that consideration into a more sophisticated characterization of goodness, so that options which satisfy the consideration will, other things equal, be better than alternatives.

Satisficing versus Maximizing

It is widely maintained that self-interested rationality is a matter of maximizing one's own good or well-being. Rationality more generally is also frequently characterized in maximizing terms: The rational thing to do in any decision context is whatever is best in terms of one's interests or will lead to the greatest preference-satisfaction. This orthodoxy is surprising given that we often justify what we do, to ourselves and others, by saying that the option we prefer is not the best but is instead simply adequate, sufficient, or “good enough,” whereas a rejected alternative “more than enough,” “more than is needed,” or simply unnecessary because we are “satisfied,” “content,” or “fine as we are.” The prevalence of such locutions inspires some, most notably Michael Slote, to propose a “satisficing” alternative to the orthodoxy, according to which a person is rational if the preferred option is good enough, where options other than the best are good enough. In the self-interested case, which will be my exclusive concern here, the satisficing claim is that it is rational to prefer an option so long as it is good enough in terms of one's own good or well-being.

Proponents of maximizing conceptions of rational choice must claim that the justifications we give in terms of adequacy or sufficiency are either illegitimate or nonliteral: “Good enough” and other similar locutions are part of legitimate justifications according to the maximizing conception only when they are used elliptically, as a way of indicating that despite appearances the course of action is in fact best.

Two Kinds of Strategies

Suppose I need to decide whether to go off to fight for a cause in which I deeply believe, or stay home with a family that needs me and that I deeply love. What should I do? My friends say I should determine the possible outcomes of the two proposed courses of action, assign probabilities and numerical utilities to each possibility, multiply through, and then choose whichever alternative has the highest number.

My friends are wrong. Their proposal would be plausible in games of chance where information on probabilities and monetarily denominated utilities is readily available. In the present case, however, I can only guess at the possible outcomes of either course of action. Neither do I know their probabilities. Nor do I know how to gauge their utilities. The strategy of maximizing expected utility is out of the question, for employing it requires information I do not have.

Nevertheless, my friends have not given up trying to help, so they point out that I could simulate the process of maximizing expected utility by assuming a set of possible outcomes, estimating their probabilities, and then making educated guesses about how much utility they would have. I could indeed do this, but I decide not to, for it occurs to me that I have no reason to trust the formula for maximizing expected utility when I have nothing but question marks to plug into it.