III. NORMATIVE ETHICAL THEORIES

The concept of normative ethics was invented early in the twentieth century to stand in contrast to the concept of metaethics. In ethical theories prior to the twentieth century, it is impossible to discern any sharp distinction between what have come to be called metaethics and normative ethics. In the first half of the twentieth century, however, this distinction began to structure ethics as an intellectual discipline and it continues to be influential at the end of the twentieth century even though crucial theoretical supports for it have disappeared.

Normative ethics was regarded as that branch of ethical inquiry that considered general ethical questions whose answers had some relatively direct bearing on practice. The answers had to be general rather than particular in order to distinguish normative ethics from casuistry; they had to have a bearing on practice in order to distinguish normative ethics from metaethics. Casuistry was understood in its classical sense as the study of particular cases, while metaethics was understood originally as the inquiry into the semantics of ethical language.

G. E. Moore's classic proposal for the structure of ethics distinguished three key questions: (1) What particular things are good? (2) What kinds of things are good? and (3) What is the meaning of "good"? The first question is the central question of casuistry, while the second question falls within normative ethics, and the third, within metaethics (although Moore used neither the term "metaethics" or "normative ethics" in his early work). Normative ethics as a field of inquiry, then, is positioned somewhat precariously between the detail of casuistry and the abstractness of metaethics.

The character of normative ethics was also strongly influenced in the first half of the twentieth century by the almost universal acceptance of the principle of moral neutrality. This principle, accepted by virtually all mainstream Anglo-American moral philosophers from the 1930s to the 1960s, asserted that the results of metaethical investigations were logically independent of normative ethics. When coupled with the original understanding of metaethics as an account of the meaning of key ethical terms, it implied that such semantic investigations were logically irrelevant to inquiries about how to live. Under the influence of this principle, normative ethics was largely abandoned by Anglo-American moral philosophers in favor of a single-minded pursuit of metaethical inquiry. And since the metaethical views most in favor during this period were various forms of noncognitivism (e.g., emotivism and prescriptivism), it was regularly asserted that normative ethics should be relegated to preachers, novelists, and other nonphilosophers. The widely accepted noncognitivist views held that there was no cognitive content to normative ethical judgments since these judgments were primarily expressions of attitudes (as emotivists held) or primarily expressions of prescriptions (as prescriptivists held). But if normative judgments had no cognitive content—if, that is, they were primarily the expression of noncognitive attitudes or imperatives—then it was unclear why moral philosophers should be concerned with examining them. Normative ethics was regarded as largely a matter of exhortation and was removed from the standard repertoire of strictly philosophical concerns.

This sharp distinction between metaethical and normative inquiry, however, together with the relegation of normative ethics to nonphilosophical inquiry, was too unstable to last. Philosophers increasingly recognized that the principle of moral neutrality was not a theoretically neutral presupposition of ethical inquiry but rather drew a considerable amount of its support from the prevailing noncognitivist view. When these noncognitivist views were severely challenged in the late 1950s and 1960s (by, among others, Philippa Foot, Kurt Baier, Stephen Toulmin, and Alan Gewirth), the sharp distinction between metaethics and normative ethics was blunted; this opened the way to a resurgence of interest in normative ethics, expressed by new attempts to reformulate and to defend classical ethical views. Although a complete historical explanation of the remarkably sudden return of philosophers in the 1960s and 1970s to the classical questions of normative theory will no doubt be extremely complex, the decline of noncognitivism and the concomitant rejection of a sharp distinction between normative ethics and metaethics surely contributed to it. Classical Kantian theory was developed in a creative and persuasive manner by John Rawls and his student, Thomas Nagel, along with Alan Donagan, Alan Gewirth, and others. Utilitarianism received new attention from, among others, Richard Hare and his students Derek Parfit and Peter Singer. The classical Aristotelian/Thomist view was reformulated and defended by Elizabeth Anscombe, Peter Geach, Alasdair MacIntyre, and like-minded moral philosophers.

What was revived under the label "normative ethics," however, was not identical to what had previously been neglected by moral philosophers as normative ethics. The watershed in ethical theory in the 1960s changed not only the interests of moral philosophers but also changed their conception of their discipline. The task of metaethics was expanded from the narrow one of clarifying the semantics of ethical terms to a much broader investigation of the whole range of metaphysical, epistemological, and semantic questions associated with ethical inquiry. Metaethics came to be concerned not only with questions about the meaning of ethical terms and judgments, but also with metaphysical questions about the nature of ethical properties and epistemological questions about how claims to ethical knowledge are to be appraised. Normative ethics in turn came to be understood as that pole of ethical theory that stood closest to practice. Whereas previously the distinction that most clearly structured ethical inquiry was the distinction between metaethics and normative ethics, the crucial distinction increasingly came to be that between ethical theory and applied ethics.

Ethical theory was distinguished from applied ethics by being both more general and more abstract, and also by being less driven by a concern that its results would have some immediate consequences for action or policy. Within ethical theory, however, elements coexisted that, according to earlier views, would have been sharply distinguished as metaethical and normative. Ethical theory inquired into the epistemological and metaphysical features of ethics as well as into the most general truths about how we should live. Also, the new conception of ethical theory held that these two kinds of inquiry were continuous; it was not possible to pursue either kind without attending to its implications for the other. Ethical theory had become a seamless web with areas of greater or less practical relevance, roughly corresponding to those areas earlier distinguished as the normative and the metaethical.

One consequence of these complex historical developments is that it has become much more difficult to give a precise characterization of normative ethics than it would have been at an earlier time. Nevertheless, certain common assumptions about the nature of normative ethics, as well as a widely shared taxonomy of the varieties of normative theory, have persisted through these developments in the concept of normative ethics. The common assumptions include the claim that the central task of normative ethics is to define and to defend an adequate theory for guiding conduct. The received taxonomy divides normative theories into three basic types: virtue theories, deontological theories, and consequentialist theories. The following section will examine these three types of normative theory with the aim of exploring their distinctive features.

Types of Normative Theory

The basis for distinguishing the three types of normative theory lies in three universal features of human actions. This recourse to the features of actions should not be surprising, since the aim of normative theory is to guide action. Every human action involves (1) an agent who performs (2) some action that has (3) particular consequences. These three features may be set out as follows:

If Jones tells a lie to Smith that causes Smith to miss his train, then Jones is the agent, his telling a lie is the action, and Smith's missing the train is one of the consequences of the action. Difficulties arise, of course, in many cases in determining whether someone is an agent in a particular case (e.g., if Jones is insane when he shoots the president, is he really the agent of any action?); or the nature of the particular action performed (e.g., if Jones is cutting down a tree, believing reasonably that he is the only one in the forest, but Smith wanders by and the tree falls on him, causing his death, does a killing take place or merely a death?); or what the consequences of a particular action may be (e.g., if Jones tells Smith "Take the stuff," but Smith understands him to say "Take the snuff," with the consequence that he takes the snuff and due to a hitherto undiscovered allergy becomes ill, is his illness a consequence of Jones's action in saying "Take the stuff"?). These are difficult questions, of course, and they have been much discussed in contemporary action theory in philosophy. In the typical case of human action, however, agent, action, and consequences can be identified, and the typical case provides the basis for the widely shared taxonomy of normative theories.

Ethical or broadly evaluative judgments can also be classified using a taxonomy drawing on these features of human action. Some ethical judgments are primarily evaluations of agents, such as "Jones is a compassionate doctor" or "Smith is a conscientious nurse." In these cases the object evaluated is a particular person, and he or she is evaluated as a possible or actual agent of an action. Some other ethical judgments are primarily about actions in the narrow sense, such as "Jones has a duty to tell the patient the truth about the diagnosis" or "The direct killing of the innocent is always wrong." In these cases, the primary object of ethical evaluation is an action—the thing done or to be done. This action may be characterized either as required ("X MUST be done") or as permitted ("X WOULD be right to do") or as forbidden ("A would be wrong to do"). More concrete characterizations of actions are also possible, such as "X WAS a vicious action" or "X WAS a heroic action." In all of the cases, however, the action is the primary object of evaluation.

A third class of ethical judgments is primarily about states of affairs or objects that are neither agents nor actions, such as "Health is more important than money" or "Human suffering is a terrible thing." Ethical judgments like these do not, directly at least, evaluate either agents or actions. However, the objects evaluated in them, may be, and frequently are, the possible consequences of actions. Thus, this last class of judgments can also be matched to one of the three basic features of human action.

Normative theories may have any of three basic structures, and the differences among these structures are determined by which of the three kinds of practical judgments is taken as basic by a particular theory. Virtue theories take judgments of agents or persons as most basic; deontological theories take judgments of actions as most basic; and consequentialist theories take judgments of the possible consequences of an action as more basic. The sense in which a theory takes a judgment of a certain kind as most basic will become clear in the discussion of each type of theory.

VIRTUE THEORIES. Normative theories that regard judgments of agents or of character as most basic are called virtue theories because of the central role played in them by the notion of a virtue. In the context of these theories, a virtue is understood as a state of a thing "in virtue of which" it performs well or appropriately. In this broad understanding of virtue not only human beings possess virtues but also certain inanimate objects—a virtue of a knife, for example, will be a sharp blade. Indeed, anything that can be said to have a function or role attached to it because of the kind of thing it is may be said to possess virtues, at least potentially.

A virtue theory takes judgments of character or of agents as basic in that it regards the fundamental task of normative theory as depicting an ideal of human character. The ethical task of each person, correspondingly, is to become a person who has certain dispositions to respond in a characteristic way to situations in the world. Differences among persons may be of quite different kinds. Some people are shorter or fatter than others, some more or less intelligent, some better or worse at particular tasks, and some more courageous, just, or honest than others. These differences can be classified in various ways: physical versus mental differences, differences in ability versus differences in performance, and so on. Those features of human beings on which virtue theories concentrate in depicting the ideal human being are states of character. Such theories typically issue in a list of virtues for human beings. These virtues are states of character that human beings must possess if they are to be successful as human beings.

Typically, a virtue theory has three goals:

1. to develop and to defend some conception of the ideal person
2. to develop and to defend some list of virtues necessary for being a person of that type
3. to defend some view of how persons can come to possess the appropriate virtues.

Virtually all ancient moral philosophers developed normative ethical theories of this sort. The ethical theories of Plato and Aristotle, in particular, provide models of this kind of normative ethical theory. As a consequence, the particular disputes that occurred among ancient philosophers centered on questions that one would expect to arise within a virtue perspective. What are human virtues? How are they acquired? Are they essentially states of knowledge? Can one know that a certain trait of character is a virtue without possessing it? Is it possible to have one, or a few, of the virtues without possessing all of them? Are all human virtues of the same type or are there fundamentally different kinds? Are human virtues a matter of nature or of convention? And, most important of all, what is the correct list of moral virtues? Much of the discussion of ethics in ancient Greece centered on a particular short list of virtues—justice, temperance, courage, and wisdom—that came to be called the cardinal virtues. After the introduction of Christianity into Europe, these four virtues were joined by faith, hope, and charity—the so-called Christian virtues—to form the seven virtues; these, together with the seven deadly vices, dominated medieval thinking about ethics.

One can also see how questions of human character are basic according to virtue theories by seeing how questions about (1) which actions one ought to perform and (2) which consequences one ought to bring about are subordinated to questions of human character. For a virtue theory the question "Which actions ought one to perform?" receives the response "Those actions that would be performed by a perfectly virtuous agent." Similarly, those states of affairs one is required to bring about in the world as a consequence of one's actions are those states of affairs valued by a perfectly virtuous person. Of course, particular actions may also be required by one's particular virtues. For example, someone who possesses the virtue of honesty may be required by the virtue itself to tell the truth in certain cases. Or someone may be required to pursue certain consequences by certain virtues. For example, an agent who has the virtue of benevolence may be required to pursue the happiness or well-being of others. But these requirements are derivative from the virtues, and the fundamental ethical question thus remains a question about the correct set of virtues for human beings.

DEONTOLOGICAL THEORIES. Deontological normative theories take moral judgments of action as basic, and they regard the fundamental ethical task for persons as one of doing the right thing—or, perhaps more commonly, of avoiding doing the wrong thing. While virtue theories guide action by producing a picture of ideal human character and a list of virtues constitutive of that character, deontological theories characteristically guide action with a set of moral principles or moral rules. These rules may refer to particular circumstances and have the following form:

Actions of type T are never (always) to be performed in circumstance C.

Or, they may be absolute in that they forbid certain actions in all circumstances and have the following form:

Actions of type T´ are never to be performed.

The essential task of a deontological theory, then, is twofold:

1. to formulate and to defend a particular set of moral rules
2. to develop and to defend some method of determining what to do when the relevant moral rules come into conflict.

One must qualify, however, the claim that deontological theories make rules fundamental in ethics. What is fundamental, in fact, are actions themselves and their moral properties. This emphasis on actions can take either of two forms: A normative theory may guide action by requiring agents to perform certain kinds of action that can be specified by a rule or other general action guide. Alternatively, one might regard normative theories as requiring particular actions that in their "particularity" elude specification by a rule. This difference has led some moral philosophers to distinguish two forms of deontological normative theories: rule deontological theories, which guide action in the first manner, and act deontological theories, which guide action in the second. Virtually all influential deontological theories, however, have taken a rule form and, for this reason, this discussion will continue to emphasize the centrality of rules.

Just as a virtue theory subordinates judgments of actions and consequences in a characteristic way, a deontological theory subordinates judgments of character and consequence. The state of character ethically most important in a deontological view is conscientiousness—that state of character that disposes persons to follow rules punctiliously, whatever the temptations may be to make an exception in a particular case. Conscientiousness does not have value in itself, but it has value derivatively because it is the most important state of character for ensuring that persons follow rules and, hence, that they do what is right. In a similar way, the consequences of actions that deontologists are most concerned with are the consequences of particular rulefollowings. Not all of an agent's practical life, however, need be reduced to rule-following. An agent may have certain personal ideals or particular projects that exist apart from moral rules. These personal ideals or personal projects may be pursued, according to the deontologist, but their pursuit is permitted only if it does not violate the moral rules. Moral rules define the limits of practical pursuits and projects. They are the moral framework within which nonmoral matters can go on. And this is the sense in which moral rules with their emphasis on judgments of actions are basic, according to the deontological view.

Just as virtue theory has its historical roots in the moral philosophy of ancient Greece, deontological theories have affinities with legalistic modes of thought characteristic of Judaic and later Roman thought. The Decalogue (Ten Commandments), although it functions in a religious context, provides a model of a set of rules of conduct that are basic in much the same way rules function in a deontological theory. One is required to follow the rules in the Decalogue because they are the commandments of God, and reasons can be given why it is appropriate to do what God tells one to do. When a deontological theory is deployed in a secular context, however, this reason for rule-following is necessarily absent. Nor can deontologists require that rules be followed because doing so is necessary to become persons of a certain sort or because doing so is necessary to bring about certain consequences. If they took the first route, their view would become a virtue theory; if they took the second route it would become a consequentialist theory. For a view to be genuinely deontological, it must claim that an agent's fundamental ethical task is to perform certain actions and that the value of this task cannot be dependent on the value of either virtues or consequences.

The most profound attempt to defend this view was anticipated in ancient moral philosophy by the Stoics and was developed in its most persuasive form by the modern German philosopher Immanuel Kant. The Stoics claimed that moral rules are expressions in the human realm of laws of nature and that rational creatures are required to follow these rules because, as creatures, they are parts of nature and, as such, obligated to bring their action in line with natural forces. Human beings differ from other objects of nature by possessing both freedom and reason. Since they are free, they may act against nature; since they have reason, however, they can understand natural laws and choose to bring their action in line with such forces. Kant's view agrees with the Stoic view in broad outline, but he develops the notions of freedom and reason far beyond the Stoic view. Kant's ultimate answer to questions about how we discover the correct set of moral rules is that only by following the dictates of reason can we be genuinely free.

CONSEQUENTIALIST THEORIES. Consequentialist normative theories take judgments of the value of the consequences of actions as most basic. According to these theories, one's crucial ethical task is to act so that one will bring about as much as possible of whatever the theory designates as most valuable. If a particular consequentialist theory designates, for example, that pleasure is the only thing valuable in itself, then one should act so as to bring about as much pleasure as possible. The goals of a consequentialist theory itself are threefold:

1. to specify and to defend some thing or list of things that are good in themselves
2. to provide some technique for measuring and comparing quantities of these intrinsically good things
3. to defend some practical policy for those cases where one is unable to determine which of a number of alternative actions will maximize the good thing or things.

Like deontological theories, consequentialist theories can be divided into act and rule varieties. Act consequentialism requires agents to perform the particular action that in a particular situation is most likely to maximize good consequences. Rule consequentialism requires agents to follow those moral rules the observance of which will maximize good consequences. The difference between these two forms of consequentialism, however, is not as straightforward as it may at first seem. It is particularly difficult to precisely characterize rule consequentialism. Is the agent supposed to follow those rules that, if followed by everyone, would maximize good consequences, or rather those rules that will maximize goodness, regardless of how other agents act? There are a number of similar difficulties in characterizing rule consequentialism, and these difficulties have led some moral philosophers to deny that there is a genuine distinction here at all. They have argued, indeed, that when any form of rule consequentialism is rigorously characterized it will be found to degenerate into a form of act consequentialism.

For consequentialists, the distinction between instrumentally good things and intrinsically good things is also of special importance. Instrumentally good things are good only insofar as they play some role in bringing about intrinsically good things. If, in a particular case, something that is ordinarily instrumentally good does not stand in the appropriate relation to an intrinsically good object, then its goodness evaporates. Its goodness is merely dependent. Intrinsically good things, on the contrary, are good not because of any relation in which they may stand to other things. Their goodness is independent because it is constituted by the kind of thing the good thing is. Thus, a particular consequentialist theory may hold that only pleasure is intrinsically good, but that other things, including types of action and states of character, are instrumentally good. The virtue of honesty, for example, might be regarded as instrumentally good by such a theory since honesty is likely to contribute to maximizing human happiness. Even if honesty is typically instrumentally good, however, situations may arise in which one could maximize pleasure by acting deviously rather than honestly. In such cases, a consequentialist theory (complications about rule versions of the theory aside) would hold that one should perform the devious action. According to this view, there is nothing about honesty in itself that is good.

Consequentialist theories find their fullest expression in modern thought, especially in the thought of the British utilitarians Jeremy Bentham, John Stuart Mill, and Henry Sidgwick. Drawing on earlier work in the British empiricist tradition, the classic utilitarians claimed that the only intrinsically good thing is human happiness, which they understood as constituted by pleasure and the absence of pain. The utilitarian maxim, "Act always in such a way as to promote the greatest happiness to the greatest number," has been the paradigmatic consequentialist moral principle and has inspired many more recent consequentialists.

There was much disagreement among classical utilitarians, however, about the details of their view. Can pleasures be distinguished qualitatively as well as quantitatively? What role should rules and virtues play within the practical thought of a utilitarian? How can the flavor of the absolute prohibitions associated with justice and the inviolability of the person be preserved within a utilitarian framework? These questions, along with other similar ones, were answered differently by different utilitarians. They were at one, however, in aspiring to formulate and defend a particular version of consequentialism.

The distinction above between the instrumentally and intrinsically good makes it possible to specify more clearly what a consequentialist theory is and to overcome certain difficulties of definition that may creep in. If a consequentialist theory is characterized as one that specifies some object, state of affairs, or property that should be maximized, one might ask whether the object or state of affairs referred to in this definition might be either a state of character or the performance of certain actions. If so, then the distinctions between a consequentialist theory, on the one hand, and a deontological theory or a virtue theory, on the other, seems to be in jeopardy. If the intrinsically valuable things specified by a consequentialist theory can include actions or states of character, then virtue theories and deontological theories would seem to be mere species of consequentialism, distinguished from other forms of consequentialism by the type of thing they specify as intrinsically valuable. Virtue theories would be consequentialist theories that specify states of character as intrinsically valuable; deontological theories would be consequentialist theories that specify the performance of certain actions as valuable. If deontological and virtue theories are merely varieties of consequentialism, however, there are not three basic structures but rather one basic structure with a number of varieties.

One might deal with this difficulty by defining a consequentialist theory as one that specifies what is intrinsically good but includes neither states of affairs nor actions, but this seems arbitrary. In addition, although this solution no longer allows that deontological theories and virtue theories are varieties of consequentialism, it does not make it possible to understand how these three types of theory exhibit different structures. One can see that there are different structures here, however, by looking more closely at the differences among these theories. Suppose that a particular consequentialist theory specifies certain virtues as the only intrinsically valuable things. Suppose, more specifically, that a particular consequentialist theory, C, specifies that the virtue of justice is the only intrinsically valuable thing. One can also suppose that a virtue theory, V, specifies the good for human beings such that it is constituted solely by the virtue of justice. Are these two theories practically equivalent? If virtue theories are a mere variety of consequentialism, they should be. If they are not, then virtue theories are not a mere variety of consequentialist theory.

One can see that these two theories are not practically equivalent by considering the practical requirements each imposes on an agent. C requires that an agent act in such a way that he or she will maximize the number of just persons. Since consequentialist theories require that agents maximize whatever is intrinsically valuable, and since the only intrinsically valuable thing according to C is the virtue of justice, agents are required by this theory to maximize justice. V, however, need not have this consequence. What V requires of an agent is that he or she develop those virtues that are constitutive of being a good human being. V requires, then, merely that an agent develop justice. There is nothing in V itself that requires an agent to try to bring about justness in others. A virtue theory more complicated than V may include a virtue—perhaps benevolence—that requires agents to promote the well-being of others as well as themselves. But this requirement to maximize the number of people who possess virtues is not a requirement derived from the nature of a virtue theory itself. It can be derived only from some particular virtue that may—or may not—be a component of a particular virtue theory.

One can arrive at this same point by considering an agent who finds herself in a situation where she can maximize the number of just persons only by becoming herself unjust. In order to make others just, she must become unjust. One example of such a case might be a politician who believes that the best way to make the citizens of her country just is to acquire political power and to exercise it in ways that only she can succeed in doing. Also, suppose she knows that only by renouncing justice herself, by being prepared to act unjustly, can she acquire political power. Thus it is only by becoming unjust that she can most efficiently make others just.

What do C and V have to say to this agent? It is clear that C would approve the renunciation of justice on her part if that would maximize the number of persons who possess justice. The loss of this particular agent's own justice to the sum of justice in the world is more than offset by the gain in the number of persons who are just. The sacrifice is worth it. But what would V require? It is equally clear that V does not require the agent to sacrifice her own justice. Virtue theories hold that an agent's own character plays a special role in his or her practical thinking that it does not play in a consequentialist theory. A virtue theory gives agents reasons to act because it is supposed that each person wants to be a flourishing and fulfilled human being. An agent's own life and character then will have a certain primacy according to a virtue theory. Virtues are not just intrinsically valuable things that should be inculcated in as many agents as possible. They are states of character that each agent must acquire in order to succeed as a human being. Thus, V will not necessarily require that this agent become unjust even if this would maximize the amount of justice in the world.

Similar conclusions follow with regard to a comparison between consequentialist theories and deontological theories. Consider a particular consequentialist teleological theory, C', that specifies that the only intrinsically valuable things are acts of truth-telling, and a particular deontological theory, D, that specifies that the only moral rule is one that enjoins truth-telling in all cases. Are these two theories practically equivalent? Again it is useful to consider a case in which maximizing a particular good requires the renunciation of it by an agent. Suppose that an agent finds himself in a situation in which he can most efficiently produce the maximum ratio of truth-tellings to lyings by himself telling a lie. Perhaps he has discovered that, by telling others that whenever they tell a lie their life is shortened by three weeks, he can most efficiently promote truth-telling. But he also knows that this is a lie. What should he do?

It seems clear that C' would require him to act in whatever way will maximize the number of truth-tellings, and, if this requires him to lie, so be it. Although his lie may be intrinsically bad, its badness will be more than outweighed by the intrinsically good states of affairs it brings about. The person who accepts D, however, believes that there is a moral rule enjoining everyone always to tell the truth. This rule gives him a reason to act, because he is committed to doing the right thing. He is not committed primarily to bringing about as many right or dutiful actions as possible; rather, he is committed to doing the right thing. Just as a virtue theory holds that an agent stands in a more intimate relation to his own character than he does to the characters of other persons, a deontological theory holds that an agent stands in a more intimate relation to his own actions than he does to the actions of others. The action of an agent who follows a moral rule will have a different moral significance for a deontologist than the action of an agent who brings it about that someone else follows a moral rule. For a deontologist, it is not as important that there be rulefollowings as that he or she follow moral rules. D need not then require, or even permit, that the agent tell a lie if this is necessary to maximize truth-telling, and hence C' and D, like C and V, are not practically equivalent. If they are not practically equivalent, however, then deontological normative theories, like virtue theories, are not mere varieties of consequentialism.

Deeper Differences among Normative Theories

This comparison of virtue, deontological, and consequentialist normative theories suggests that the differences among them are deeper than might at first appear. Indeed it suggests that while they certainly differ with regard to which of the three kinds of practical judgments they take as most basic, there are other, and more fundamental, differences among them. To accept one of these normative theories is to accept a particular attitude toward the relation of an agent to his or her character and actions. If one adopts a virtue theory, one's own character comes to have an especially important place in one's practical thinking. It is of the first importance that one become a person of a certain sort. This view need not imply, as it may seem to, that one is committed to an egoistic or selfish life. One may be guided by a virtue theory to pursue a life dominated by generosity and concern for others. One may, indeed, strive to become completely selfless in the sense of always putting the needs of others ahead of one's own needs. But even if this is one's goal, it is also true that one's own character forms the primary focus of one's practical life. The apparent combination here of concern for self and concern for others may appear paradoxical, but it is surely not incoherent. Some of the greatest moral heroes—for example, Gandhi, Jesus, and Albert Schweitzer—seem to have combined these two concerns in their lives.

In a similar way, if one adopts a deontological theory, one's own actions come to play an especially important role in one's practical thinking. It makes a difference to one that one's actions are wrong. It is more important practically to an agent that he or she has told a lie than that a lie has been told. In cases where one's telling a single lie will prevent three others from telling lies, one will not decide what to do by simple arithmetic. Of course, a deontologist will not expect that others will have the same concern for her lie as she will have for it. She may recognize that for someone else, his telling a lie will have a different practical significance for him than her telling a lie will have for him. And just as she may not be prepared to tell one lie to prevent him from telling two, she will not expect him to tell one lie to prevent her from telling two. Indeed, she will recognize that from his point of view, his telling one lie is worse in an important sense than her telling two, just as from her point of view her telling one lie is worse than his telling two.

The special significance given to one's actions by a deontological theory need not imply that a deontologist is egoistic or, in the ordinary sense of the term, self-centered. In this way the deontologist is in a situation similar to that of the virtue theorist. The particular moral rules that one is required to follow may give the needs and interests of others parity with one's own, or, more likely, they may require one to put others ahead of oneself. What they cannot require is that one take up a particular attitude toward the rules themselves. The rules cannot, as it were, define their own condition of application—nor can they specify how they relate to one's faculty of practical decision making at the deepest level.

To a consequentialist, giving this special significance to one's character or one's actions may seem confused and possibly morally corrupt. Of course, consequentialists may be concerned with questions of character, but character cannot be their central normative focus. According to consequentialism, what is of primary ethical importance is that the amount of the intrinsically valuable be maximized. Determining the most effective means for maximization involves straightforward questions of efficiency. These questions may be neither simple nor easily answered, but structurally they are straightforward: Which of the possible courses of action will most likely maximize the amount of goodness in the world? In canvassing the possible means to this end, the consequentialist requires an agent to throw his own character and actions into the same category with other possible means. The kind of character one should develop depends upon the kind of character that will contribute most to the relevant goal. The actions one should perform depend similarly on consequentialist goals. For a consequentialist, one must put a certain distance between oneself—considered as the agent who must make practical choices—and one's own character and actions. One's character and actions have the same role in one's practical thinking as would any other possible means—one's wealth, for example, or influence—that are in a more usual sense external. More important, one's own character and actions have no more special role in practical thinking than do the character and actions of others. All are regarded as possible means to maximize intrinsically good things, and one's own actions and character may have special significance only insofar as they may be more easily—because more directly—manipulated by oneself.

One might think, however, that one feature of the agent's character cannot be treated as a mere means, even by a consequentialist. For any consequentialist theory, it will surely be important that persons have those states of character that dispose them to pursue or to favor intrinsically good things. It might be argued that this state of character cannot be treated by the theory as a mere means. But this argument underestimates the resources within consequentialism for distancing an agent from his or her character. Suppose an agent holds a consequentialist normative theory, C', according to which the only intrinsically good things are states of human pleasure. Suppose also that this agent has a character such that he is disposed always to act in ways he believes will maximize human pleasure. This argument suggests that this agent will not be prepared to sacrifice for the goal of maximal pleasure his own disposition to pursue this goal. But why should this be the case? One might think that a case could never arise in which an agent could contribute most to maximizing pleasure by changing his character to that of someone unconcerned with maximizing pleasure. But this view is surely wrong. Suppose the agent discovers an empirical law according to which human pleasure is maximized only if agents are disposed not to pursue human pleasure but to pursue knowledge. But if this is true—and it is surely possibly true—the agent should act to change as many persons' characters as possible from pleasure-seeking to knowledge-seeking characters. Nor is there any reason why, on consequentialist grounds, this agent should make an exception in his or her own case. So even those features of human character that lead an agent to pursue the maximization of intrinsically good things are not given a special place by consequentialists. Every feature of the character of an agent may be regarded as a possible means to the maximization of the relevant goal.

This feature of consequentialist theories was first emphasized by Henry Sidgwick, the greatest of modern utilitarians. Sidgwick was convinced that if the utilitarian goal of human happiness was to be maximized, then it was necessary that most persons not be utilitarians. Indeed, he thought that what was probably required was that most persons hold deontological views and have their character shaped in accordance with such views. He proposed then, for utilitarian reasons, that utilitarianism be propagated as an esoteric view, and that only a few of the most able and intelligent members of society have their characters shaped in accord with it. These bearers of the esoteric view, in turn, would mold the characters of those less able and enlightened in accord with a deontological perspective. Had Sidgwick's enlightened few become convinced that maximal human happiness required that they, too, acquire "deontological characters," simple consistency would have required them to change their own characters appropriately. In this way, consequentialism might require that agents strive to bring about a world in which no one, not even oneself, has the kind of character that would dispose one to strive at the most basic practical level for consequentialist goods.

Justifying Normative Theories

The question of how, if at all, one can rationally choose among these three normative theories is a question taken up under the topic of moral epistemology. It is important to note here, however, that these normative theories emerge in Western thought as components in comprehensive philosophical theories developed by Plato, Aristotle, Aquinas, Kant, Mill, and other major philosophers. They are embedded in rich and complex worldviews in ways that make it difficult to discuss them in isolation from their theoretical and historical settings.

The tendency within contemporary ethical theory is to discuss the merits of these views in purely ethical terms and to ignore to a large extent their larger theoretical settings. Thus, consequentialism is frequently attacked because it is alleged to countenance the judicial punishment of the innocent if that is required for achieving some good end. In arguments like this one, the alleged ethical implications of a normative theory are appealed to in order to evaluate the theory. Similarly, deontologists may be criticized for holding that certain actions are morally forbidden even if performing them in a particular case might prevent an enormous tragedy. It is now a matter of record that these arguments have been unsuccessful in producing agreement within normative ethics. Nevertheless, the same slightly tired arguments continue to be made.

The lesson from the history of these views would seem to be, however, that if any of them is to be adequately defended, or successfully criticized, its theoretical setting must be taken into account. Each of these theories has complex relations with particular philosophical accounts of rationality, explanation, nature, intention, the law, the passions, and other topics of central philosophical interest. A more adequate account of them, if possible here, would have to take these theoretical entanglements into account. Certainly any serious attempt to choose rationally among them would have to locate them in this larger theoretical setting.

Normative Ethics and Practice

The raison d'être for normative ethics, as we have seen, is to guide action, and the theories explored above have been developed with such guidance in mind. There is general disagreement, however, about exactly how these normative theories are to relate to the resolution of particular normative problems. It is not easy to demonstrate how the debate between consequentialists and deontologists is related to more concrete disagreements about physician-assisted suicide or recombinant DNA research. Part of the difficulty arises from the fact that each of the three normative theories embodies a particular conception of how it relates to concrete normative problems. There is no theory-independent criterion of how normative theories are to guide action, since each theory embodies a view about its own application. In this way normative theories double back on themselves with regard to their action-guiding function.

An illustration of this doubling-back phenomenon is found in current debates about the relation of virtue theories to practice. Virtue theories are frequently criticized because they do not yield concrete action guides in the way that consequentialist and deontological theories appear to do. The moral advice to "Be just" lacks the action-guiding bite of either a moral rule that requires an agent to perform certain actions or a consequentialist conception that specifies some good to be maximized. But this objection fails to take account of the distinctive way in which virtue theories purport to guide action. A central claim of virtue theories is that the action-guiding function of a normative theory is not to resolve concrete puzzles about action. Edmond Pincoffs, a leading contemporary virtue theorist, coined the useful term "quandary ethics" precisely to designate what virtue theories are against: a conception of normative ethics as guiding action by giving a particular solution to quandaries about action. If one supposes that the only way in which a normative theory can guide action is by resolving particular moral quandaries, then one is unlikely to take virtue theories seriously.

Virtue theories offer, however, an alternative account of the action-guiding function of normative theories. They claim that an adequate normative theory will prescribe something like a training program to make agents ethically "fit." This program may not specify exactly how one is to act in particular cases, because these decisions are best left to the prudential decisions of a "morally fit" agent in the concrete decison-making situation. Thus, virtue theories double back on themselves and specify how they are to relate to practice. Both deontological and consequentialist theories also contain such self-referential accounts of their own application.

An important implication of this doubling-back phenomenon is that one cannot assess the adequacy of normative theories by invoking a well-defined criterion for "successful" action-guiding without begging the question. To have such a well-defined criterion is already to have taken a position on some of the fundamental questions in normative ethics.

This difficulty is actually even more serious than this first point suggests. It is not just that each of the three normative theories embodies a well-defined criterion of how normative theory should relate to practice. Also, there are a number of different models of how general ethical thinking should relate to concrete practice. Some of these models have loose affinities with some of the normative theories, but there is not a fixed or necessary connection between them. Indeed, the conflicts among the normative theories cut across, in complex ways, the conflicts among these models for relating normative theory to practice. A representative collection of these models would include: (1) deductivism,(2) dialectical models, (3) principlism, (4) casuistical models, and (5) situation ethics. These models have been for the most part badly defined in the current literature, and the differences among them and their relations to traditional normative theories tend to be matters of dispute.

DEDUCTIVISM. The deductivist model regards the actionguiding function of ethical theory to be the development of highly abstract and general first principles that, together with some factual description of a particular morally problematic situation, will entail concrete action guides. According to this model, moral principles developed and defended within normative ethical theory will play the role of premises in deductive arguments for ethical judgments about particular cases. This model of application is particularly attractive to some deontologists and consequentialists. It is related to more general accounts of justification in contemporary epistemology that suggest that all justification must come from some set of foundational claims in the area in question.

It also makes large demands on the justificatory resources of a normative theory, since all of the justification for the principles must come from the theory itself. There is no "bottom up" justification from particular moral beliefs to general principles, as will be found in some of the other models.

DIALECTICAL MODELS. Partly because of worries about the foundationalist character of deductivism, some moral theorists understand the relation between normative theory and practice in a dialectical way. Instead of supposing that justification is exclusively "top down," they suppose that there is dialectical interplay between the principles in a normative theory and particular moral judgments. Normative principles may be modified if they fail to fit our deeply held particular moral beliefs, just as our particular beliefs may be modified in order to fit principles. Whether agents modify principles or particular judgments will depend upon their degree of commitment to each and to the other beliefs they might hold. Just as the deductivist model has affinities with foundationalist theories in epistemology, the dialectical model is inspired by coherentist epistemological theories, which suggest that justification in general is to be understood as a function of how large sets of propositions "hang together" or cohere. The most influential form of the dialectical model is John Rawls's "method of reflective equilibrium," which he uses to support his deontological normative theory.

PRINCIPLISM. Some philosophers have wanted to downplay the importance of normative theory for resolving concrete ethical problems. They emphasize, for example, that consequentialist and deontological normative theories in most cases mandate the same actions, and that it is only in exceptional cases that differences seem to emerge. And they add that the exceptional cases are likely to be so difficult to resolve that both consequentialists and deontologists disagree among themselves about what normative theory requires. They conclude that general ethical reflection should focus on what they call "middle-level" principles, that is, not the most general principles in any normative theory but those that are likely to be acceptable to adherents of different normative theories. They hope that agreement may be easier to achieve in practical matters if the premises for practical arguments are not sought at the deepest level of normative theory. This model has been especially influential in bioethics and has been developed and defended by Tom Beauchamp and James Childress (1989). The middle-level principles they propose are labeled autonomy, beneficence, nonmaleficence, and justice. Their claim is that these principles, when suitably refined, are likely to be acceptable to both rule consequentialists and deontologists.

CASUISTICAL MODEL. Some philosophers have understood genuinely practical and action-guiding thinking in a way that makes it even more remote from the disputes among the classical normative theories. They propose that the appropriate model for practical reflection is found in the case-based approach popular in late medieval and early modern moral thought. According to this approach, ethical reflection should focus on certain paradigm cases of morally good action or morally bad action. Arguments from these paradigm cases to more problematic cases may be made by exploring similarities and differences between the two. This approach rejects attempts to formulate the goodness or badness of paradigm cases in abstract and general principles, and emphasizes analogical as opposed to deductive reasoning. Albert Jonsen and Stephen Toulmin (1988) have been the leading advocates of this model in recent normative ethics.

SITUATION ETHICS. Some might suggest that situation ethics is not so much a model for practical thinking as a rejection of any model. It claims that one should approach the resolution of particular moral problems by eschewing all general action guides in favor of concentrated attention to the details of the particular situation. In some of its versions it may look a bit like the casuistical model; but in its most radical formulations it would mandate that even paradigm cases should play no central role in particular reflection because they could deflect the agent's attention from the particular features of the case under consideration. Among contemporary thinkers, Joseph Fletcher has been the most prominent advocate of this view, although his early commitment to situation ethics developed later into a more general commitment to consequentialism. In his formulation of situation ethics, he suggests that reflection on particular cases should be guided by the general principle, "Do the loving thing!" However, he is insistent that this principle does not play the role of a premise in any deductive practical argument.

These five models represent different ways of thinking about how ethical reflection might be brought to bear on particular moral problems. They range from deductivism, in which successful ethical reflection requires premises drawn from an adequate normative theory, to situation ethics, which eschews any dependence on normative theory. The other three theories occupy the middle ground between these two extremes. In contemporary ethics there is no consensus on which of these models is most adequate. Each has its defenders and its critics, and there is a lively discussion in the contemporary literature about their respective merits.

When this disagreement about the correct approach to concrete ethical reflection is added to the disagreement among classical normative theories, it is easy to see why contemporary applied ethics involves conflicts of such depth and complexity. One is confronted not only with competing normative theories, but also with competing conceptions of how such theories would relate to concrete ethical problems. These two different levels of disagreement indeed tend to reinforce one another, since particular disagreements at each level tend to be tied to particular disagreements at the other.

Normative Theories and Bioethics

The revival of normative ethics in the 1960s was associated with a general renewed interest, across Western culture, in applied ethics and especially in bioethics. Rational reflection on the difficult ethical issues associated with the expanded technological resources of the biological sciences demanded a theoretical structure of some richness, and the classical normative theories provided that structure.

The conflicts between deontological and consequentialist theories have been particularly salient in discussions within bioethics. Indeed, some general discussions of bioethics and many popular textbooks treat these two options as if they are the only possible theoretical perspectives. Part of the explanation for this is surely that so many of the ethical problems in medical practice, as well as in the biological sciences more generally, involve questions about whether actions that are generally regarded as morally problematic can be justified in cases where they appear to promise great benefits. Examples of this kind of conflict are plentiful in contemporary bioethics: Can information obtained by a physician in a doctor-patient encounter be revealed to a third party without the patient's consent, if doing so will prevent some great harm? Can physicians lie to their patients in cases where doing so will increase the effectiveness of therapy and decrease the chances of severe depression? Can physicians override the religious objections of patients to certain therapies when it is clear that these therapies will provide important benefits to the patients?

Moral difficulties like these have been at the center of contemporary discussions in bioethics from its inception. They lend themselves to an analysis that regards them as embodying a general conflict between the thought that some actions (e.g., revealing confidential information, lying, or paternalistic interference) are simply not to be done and the thought that one should be prepared to do whatever is necessary so that things go as well as they can. This conflict in turn seems very close to the fundamental issues at stake between the deontologist and the consequentialist.

Until recent years, virtue theories have been conspicuously absent from most discussions of bioethics. The renewed interest in these approaches is associated with their revival within moral philosophy generally. But there are also features of contemporary bioethics that explain the attention they receive. First, a kind of impasse has developed between consequentialist and deontological approaches to some bioethical problems, and bioethicists have turned to virtue theories with the hope that they can avoid this impasse. Second, there is a new interest in questions about the character of the various agents (e.g., physicians, nurses, researchers, and technicians) who work in settings where bioethical issues arise. This interest in character is partially a reflection of impatience with "quandary ethics." It also, however, grows out of the search for new models of moral education. Molding and shaping character has seemed to many a more attractive goal for moral education than the goal of inculcating rules. Shaping character indeed seems especially important in bioethics, where change is endemic and rules become outdated quickly.

Finally, virtue theories seem to be attracting more attention within bioethics because of the strong analogies between the notion of health and overall biological fitness, on the one hand, and, on the other, the more general notion of human flourishing that lies at the heart of virtue theories. For those who think that bioethical issues are best approached by getting clear on the goals of the biomedical sciences, this analogy is likely to lead them to take virtue theories seriously.

In spite of the recent revival of virtue ethics both within bioethics and within moral philosophy more generally, however, the dominant argumentative strategies in bioethics continue to be drawn from the deontological and consequentialist traditions. Nevertheless, each of the three traditions is now represented in the contemporary bioethical discussion by competent and enthusiastic advocates, and it seems certain that the central problems within bioethics will continue to be discussed in terms contributed by these normative traditions.

w. david solomon (1995)

bibliography revised

SEE ALSO: Care; Casuistry; Communitarianism and Bioethics; Contractarianism and Bioethics; Double Effect, Principle or Doctrine of; Emotions; Obligation and Supererogation; Human Rights; Natural Law; Principalism; Utilitarianism;Virtue and Character; and other Ethics subentries

BIBLIOGRAPHY

Alora, Angeles Tan, and Lumitao, Josephine M., eds. 2001. Beyond Western Bioethics: Voices from the Developing World. Washington, D.C.: Georgetown University Press.

Anscombe, G. E. M. 1958. "Modern Moral Philosophy." Philosophy 33: 1–19.

Beauchamp, Tom L., and Childress, James F. 1989. Principles of Biomedical Ethics, 3rd edition. New York: Oxford University Press.

Beauchamp, Tom L., and Walters, LeRoy, eds. 1999. Contemporary Issues in Bioethics (Fifth Edition). Belmont, CA: Wadsworth Publishing Company.

Brandt, Richard. 1979. A Theory of the Good and the Right. Oxford: Clarendon Press.

Broad, Charlie D. 1930. Five Types of Ethical Theory. London: Routledge & Kegan Paul.

Curran, Charles E. 1999. The Catholic Moral Tradition Today: A Synthesis. Washington, D.C.: Georgetown University Press.

Donagan, Alan. 1977. The Theory of Morality. Chicago: University of Chicago Press. A defense of a comprehensive deontological normative theory.

Dwyer, Judith A., ed. 1999. Vision and Values: Ethical Viewpoints in the Catholic Tradition. Washington, D.C.: Georgetown University Press.

Fletcher, Joseph. 1966. Situation Ethics: The New Morality. Philadelphia: Westminster.

Foot, Philippa. 1978. Virtues and Vices; And Other Essays in Moral Philosophy. Berkeley: University of California Press.

Frankena, William K. 1973. Ethics, 2nd edition. Englewood Cliffs, NJ: Prentice-Hall.

Gomez-Lobo, Alfonso. 2002. Morality and the Human Good: An Introduction to Natural Law Ethics. Washington, D.C.: Georgetown University Press.

Jonsen, Albert R., and Toulmin, Stephen E. 1988. The Abuse of Casuistry: A History of Moral Reasoning. Berkeley: University of California Press.

Kavanaugh, John F. 2002. Who Counts as Persons? Human Identity and the Ethics of Killing. Washington, D.C.: Georgetown University Press.

Kittay, Eva Feder, and Meyers, Diana T., eds. 1987. Women an Moral Theory. Totowa, NJ: Rowman and Littlefield.

MacIntyre, Alasdair C. 1981. After Virtue: A Study in Moral Theory. Notre Dame, IN: University of Notre Dame Press.

Macklin, Ruth. 1999. Against Relativism: Cultural Diversity and the Search for Ethical Universals in Medicine. New York: Oxford University Press.

McGee, Glenn, ed. 1999. Pragmatic Bioethics. Nashville, TN: Vanderbilt University Press.

Moore, G. E. 1903. Principia Ethica. Cambridge: Cambridg University Press.

Moser, Paul K., and Carson, Thomas L., eds. 1997. Morality and the Good Life. New York: Oxford University Press.

Nagel, Thomas. 1970. The Possibility of Altruism. Oxford: Clarendon Press.

Oderberg, David S., and Laing, Jacqueline A., eds. 1997. Huma Lives: Critical Essays on Consequentialist Bioethics. New York: Macmillan.

Parfit, Derek. 1984. Reasons and Persons. Oxford: Clarendon Press.

Pellegrino, Edmund D. 2000. "Bioethics at Century's Turn: CanNormative Ethics Be Retrieved?" Journal of Medicine and Philosophy 25(6): 655–675.

Rawls, John. 1971. A Theory of Justice. Cambridge, MA: Harvar University Press.

Ross, W. D. 1930. The Right and The Good. Oxford: Clarendon Press.

Scheffler, Samuel, ed. 1988. Consequentialism and Its Critics. Oxford: Oxford University Press.

Sidgwick, Henry. 1907. The Methods of Ethics. 7th edition. London: Macmillan.

Smart, John Jameson Carswell, and Williams, Bernard A. 1973. Utilitarianism: For and Against. Cambridge, Eng.: Cambridge University Press.

Sugarman, Jeremy, and Sulmasy, Daniel P., eds. 2001. Methods in Medical Ethics. Washington, D.C.: Georgetown University Press.

Thomasma, David C. 2000. "Medical Ethics: Its Branches and Methods." Philosophical Inquiry 22(4): 7–23.

Tong, Rosemarie P. 1993. Feminine and Feminist Ethics. Belmont, CA: Wadsworth Publishing Company.

Wildes, Kevin William. 2000. Moral Acquaintances: Methodology in Bioethics. Notre Dame, IN: University of Notre Dame Press.

Williams, Bernard A. 1985. Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University Press.