2009 Irving and Jeanne Glovin Award Winner
Introduction: Morality as Description and Instruction
A certain counterexample is often raised against Immanuel Kant’s deontological theory of morality. In the example, a moral agent is hiding a friend or relative from a group of Nazis. The agent knows the Nazis intend to harm their friend. A Nazi knocks on the agent’s door and asks to know the whereabouts of the friend. For Kant, morality requires that the agent tell the Nazi where the friend is hiding, as according to his moral theory, it is never permissible to lie (Kant 1785). This story acts as a counterexample to Kantian ethics because it shows that in at least this case Kant’s deontology conflicts with common intuitions about right and wrong. It is generally thought to be permissible to lie in some situations, especially to save a friend from harm. Common sense would say that it is morally permissible to lie to a Nazi to keep a friend safe.
As well as highlighting a possible difficulty for the Kantian moral picture, this story calls attention to the common assumption that moral theories are accountable to out moral intuitions. The ‘Nazi at the door’ example supposedly works as a refutation of Kant, by showing that his deontology does not always match up with common sense morality. (Kant brings this example up himself to illustrate how the demands of his morality sometimes conflict with intuition.) This criticism, if it is to work, assumes that a moral theory must link up somehow with our moral intuition in order to do what we want it to.
Another counterexample that uses the same strategy is often staged against the moral theory of utilitarianism. In this example, a homeless man with no friends or family is taken to a hospital. The man’s illness is easily treatable. However, another man at the hospital is terminally ill and needs a heart transplant. This other man is a beloved husband and father of four. He is also the author of many bestselling novels that bring joy to millions of readers. Utilitarianism holds an act to be right so long as it is conducive to the greater aggregate hap[pines (Mill, 1861).According to utilitarianism then, it would be morally obligatory for the doctors to take the heart of the homeless man and give it to the father/author as this action makes a greater contribution to the aggregate happiness than simply treating the homeless man. Since taking the heart of a homeless man in order to save others is generally taken to be wrong, utilitarianism seems to prescribe an action that is wrong according to common moral intuition.
Granted, these counterexamples are too simplistic. Both Kantian deontology and consequentialism, the larger subset of theories to which utilitarianism belongs, can give sophisticated responses to these examples or be modified to avoid these problems altogether. I mention them merely to illustrate how this type of criticism is supposed to work. Each story acts as a refutation of a moral theory by showing a case wherein what the theory prescribes does not match what common moral intuition prescribes. These criticisms therefore assume that in order to be correct a moral theory must be consistent with moral intuition or, at the very least, that a moral theory must be incorrect if it conflicts too much with moral intuition.
When deciding which theory best describes the underlying principles of good human conduct, a question we can reasonably ask ourselves is, what should the relationship between a moral theory and our moral intuition be? If these counterexamples work the way they are thought to then it seems that moral theory is supposed to be consistent with intuition.
However, the purpose of moral theory is not simply to capture our intuition. Another job we seem to have in mind for moral theory is that it be instructive. We are interested in what is right and wrong and sometimes our intuition is unsure which is which, or worse sometimes we may think we know but are mistaken. We want a moral theory that is able to enlighten us in cases where what is right or wrong is unclear. To borrow a phrase from Peter Schotch, the goal of moral theory is to close our moral intuitions under the laws of deduction (Schotch, personal communication). We want a moral theory that is consistent with our existing intuition, and that will allow us to correctly deduce which actions are right and which are wrong.
The traditional moral theories of deontology and consequentialism mentioned above accomplish the two fold task I have outlined by giving general principles or rules which attempt to both capture our moral intuition, and give them the means by which the rightness or wrongness of acts may be determined. In the case of Kantian deontology, an act is measured against one of three formulations of the categorical imperative (Kant, 1785). For consequentialism, an act is correct so long as it is conducive to the greater general good (Kagan, 1989). For both of these theories determining what is right or wrong is simply a matter of learning a principle and then applying it to a situation (Dworkin, 1995). This method has the advantage of being concise and easily taught.
Nevertheless, as illustrated, one is often able to give counterexamples that show cases where these theories do not match onto experience. Although not a problem in itself, as moral intuition may in some cases be incorrect. This potential for inconsistency between theory and experience is indicative of a larger problem. In order to generalize over morality, these theories try to give unifying principles that underlie all of morality. In this paper I will argue that attempts to give unifying accounts of morality is misguided. I will argue that such projects fail because they mistakenly assume that morality is a consistently unified category. Thus, no single principle will be able to capture all of morality. To illustrate, I will discuss the basic structure of the argument given by Shelly Kagan in her book The Limits of Morality. I use Kagan as an example because her work most clearly demonstrates the problem I associate with unifying theories of ethics, that is their disconnect from moral phenomena. I will however, show that the same problem confronts all principles accounts of morality including those that more accurately match onto moral intuition. I will then, drawing on the work of W.D. Ross Gerald, Dworkin, and Jonathan Dancy, make some progress toward giving a unprincipled account of morality.
Morality Against Intuition
Kagan reveals herself as a consequentialist in the beginning of her book by her terse description of morality. “Morality requires that you perform -of those acts not otherwise forbidden- that act which can be reasonably expected to lead to the best consequences overall” (Kagan 1989, p. 1). By claiming that it is the consequences of acts that are morally important, Kagan aligns herself with the consequentialist movement. However, the fact that Kagan argues for a form of consequentialism is immaterial for my discussion. What I am more interested in, is the role that moral intuition plays in her account of morality. Kagan argues that our moral intuitions should not be taken as authoritative. Instead she argues for a more demanding and counterintuitive morality.
Kagan divides the moral terrain into a spectrum characterized by three positions, the extremist, the moderate, and the minimalist. The extremist holds that the “demands of morality pervade and every aspect of our lives” (Kagan, 1989, p.2). An extremist would be morally obligated to forgo going to a movie and getting a popcorn as both the time and the money could be used in ways more conducive to the best overall consequences. For instance, it could be used toward famine relief. The moderate is the common sense moralist who holds that moral intuition is generally right. For the moderate going to a movie is just fine. However, if on the way to the movie the moderate discovers a child drowning in a local pond then she is obligated to do what she can to rescue the child. On the other end of the spectrum, the minimalist holds that even the requirements of common sense morality are too demanding. So if a moderate discovers a drowning child on the way to a movie, she is not obligated to save the child as such an endeavor might soak her clothes and maybe make her late for the movie. Kagan’s project is to defend the extremist position, by showing that the moderate is unable to defend her position without collapsing into the minimalist or extremist position.
Kagan begins by accepting consequentialism as a given. She supposes that in general an act is thought to be right so long as it leads to the best overall consequences. She claims that only the extremist consistently strives to act in accordance with the principle. She argues that although e moderate believes in “occasional requirements to promote the good” the moderate denies “the claim that the agent must – in general – do all he (permissibly) can do to promote the good” (Kagan 1989, p.9).
To explain, Kagan uses the notions of moral constraints and moral options. A moral constraint is a special moral obligation. For instance, it is generally accepted that one is required to provide care for one’s children; this obligation is a moral constraint. A moral option is an act that is morally permissible to do but nevertheless does not lead to the best overall consequences. Going to a move then is a moral option. The act of going to a movie is morally permissible even though not necessarily instrumental to the best overall consequences.
The important difference between the extremist and moderate is their respective position on moral options. The moderate believes that moral options exist, whereas the extremist does not. Kagan characterizes the moderate as holding that once all moral obligations are met, an agent is free to act as she will so long as her acts are morally permissible. The extremist, on the other hand does not believe in the existence of moral options. According to the moral extremist, one is always required to act in such a way as to promote the best overall consequences. Kagan argues that while the moderate believes in both moral constraints and options, the extremist can hold a coherent position against the existence of options with or without constraints. The extremist can allow for moral constraints so long as once they have been attended to, the agent will then act in ways that will result in the best overall consequences. However, as Kagan argues, since the moderate requires moral constraints in order to argue for the existence of moral options and the extremist does not believe in options, the burden of justification for those constraints is placed “squarely on the shoulders of the moderate.” (Kagan, 1989, p. 10)
According to Kagan, the moderate must tell a story about how moral options are permissible. In other words, she must argue that there are some cases where an agent is exempt from the demands of morality and is free “to pursue their own interests, at the possible expense of the good.” (Kagan, 1989, p. 19). Kagan claims that it will be impossible for the moderate to allow for options that allow harm, without also allowing options to do harm. She argues that any limit on the demands morality that would permit allowing harm in the way the moderate wants, would also allow options to do harm in a way the moderate does not want. Here we see how the moderate risks collapsing into the minimalist position. The moderate wishes to have space wherein they are free from the demands of morality. But if such a space exists, it seems not only to allow harm but to allow to do harm. (Kagan, 1989)
There is one possible solution for the moderate. Perhaps if she can show that certain constraints against harm do exist, she will be able to justify cases where harm is allowed but where causing harm is not. This is why the moderate depends on the existence of moral constraints in order to justify moral options. Space here prevents me from fully accounting for Kagan’s catalogue of ways a moderate might argue for moral constraints. Kagan argues, however, that all strategies the moderate can use consist of “stressing the badness” of harm (Kagan, 1989, p. 28). This emphasis she argues, will always support the extremist positions that harm in all cases must be avoided. The moderate here is in danger of collapsing into the extremist position. The moderate is therefore unable to carve out a space where they are morally exempt.
Kagan concludes that the moderate position is untenable. This is a problem, as the moderate position embodies the position of common sense morality. Kagan argues that it is time to give up moral intuition on the grounds of being inconsistent. Instead, we should adopt the much more demanding less intuitive extremist position.
Such a change in our moral thinking, Kagan claims, would have far reaching implications into our lives.
There is no limit to what you might be called upon to sacrifice in the pursuit of good. Your material possessions, time, effort, bodily parts, or life itself – all of these might be commandeered by morality, and put to purposes quite unlike those to which you would dictate were morality’s demand less severe. (Kagan, 1989. P. 21)
Kagan proposes a complete overhaul of our moral life. Morality according to Kagan permeates our every action. She argues that in order to be moral we must then let go of the idea of moral options and that we are free to pursue our personal interests. Instead, we should more diligently follow the strict principle that all our acts be toward the greater general good.
The Problem of Principle
Kagan’s argument is compelling. Indeed, given the formulation of consequentialism she espouses, morality would be as demanding as she claims. However, I think that her conclusion, that in order to be moral one must be willing to give everything up to the greater good, is obviously false. Kagan is not justified in such an extreme abandonment of moral intuition. My task then is to find out how Kagan’s argument fails. But first I will discuss why one cannot accept her conclusion.
As noted earlier, it appears that one task a moral theory must accomplish is to capture our moral intuition. It is necessary now to explain why this is the case. W.D. Ross explains in his book The Right and the Good the relevance of intuition to moral theory (Ross, 1930). Ross argues that we would not base natural science on what reasonably thoughtful people think about the objects of scientific investigation before they have studied them scientifically. Ideally we would want to study the objects directly before theorizing about them. Ross then claims,
In ethics no such appeal is possible. We have no more direct way of accessing to the facts about rightness and goodness and about what things are right and good, than by thinking about them; the moral convictions of thoughtful well educated people are the data of ethics. (Ross, 1930, p. 40)
Although the merit of Ross’s views on science can be questioned, his explanation points out something interesting about the nature of morality. That is, that ethics has at its origin in our intuitions about what is right and wrong. Had we no such intuitions we would have no need for moral theory. Any account of morality that concludes that moral intuition be dismissed denies the very data that moral theory set out to explain.
Kagan’s conclusion that we should trade in moral intuition for a more systematic form of consequentialism must be, according to Ross’s account, false. Although Kagan’s argument is consistent and her conclusions do follow, by abandoning moral intuition Kagan’s theory no longer describes morality but something else entirely. She has given a theory that shows the implications of the imperative ‘to act always in ways that promote the good.’ But she has given no reason to accept that this imperative describes what it means to be moral. One could imagine a consistent theory wherein it is an imperative that one always act in ways to produce the most kidney beans. One could then work out all the implications that follow from such a theory, but it would be unclear how that theory would be relevant. Likewise, unless Kagan can show why the general good is the end to which morality seeks or indeed that morality seeks ends at all, it is not clear how her theory can attach itself to what we call morality. Her theory, if it abandons intuition, must be about something else.
Since Kagan’s argument is valid and her conclusion false she must be mistaken about one of her assumptions. As outlined earlier, as well as capturing out intuitions, a moral theory is supposed to enlighten us. Kagan makes this point by claiming that a moral theory must give more information then that is contained in the theory (Kagan, 1989, p. 12). A moral theory is supposed to tell us what is right and what is wrong in cases where we are unsure or mistaken.
It is understandable why we would want a moral theory to be enlightening in this way. Most of us have an interest in promoting moral behavior, whatever that comes to. However we are not always sure what the right thing to do is in some cases. For example, one could ask oneself whether it is permissible to download movies or music off the internet without paying the relevant royalties. More seriously perhaps, one could wonder whether one is within their right to have an abortion. A correct moral theory is desired to give instruction in such situations. Furthermore, we are aware of many moral atrocities that were only recognized as such until later. For example, black slavery in North America may not have been thought to be morally problematic by many slave owners at the time. Also, the sexual harassment of women in the workplace ha not until only recently been recognized as wrong (Fricker, 2006). One hope for moral theory is to give us the means to recognize situations like these so that we will be able to see ways that we might be mistaken about our moral judgements.
Given this motivation, theorists like Kagan tend to ignore the requirement that moral theory be linked to intuition in a real way. Kagan and others want to give a theory that is able to clearly designate the right and the wrong in all cases. In order to accomplish this task, Kagan explains that our theory must be adequately suited. “In moral philosophy we want to apply the same sort of criteria that we use for theory building generally: we want our moral theory to have simplicity, power, and coherence” (Kagan, 1989, p. 11). Implicit in this claim however, is a tendency to model moral theory after scientific theory. In her quest for simplicity, power and coherence, Kagan has set herself up to search for a unifying principle of morality, a law of gravity for ethics. Kagan’s desire for a unifying theory of ethics can be seen as inspired by advances in physics where certain phenomenon are discovered to be manifestations of other phenomena. For example, when physicists discovered that electric phenomena and magnetic phenomena could both be explained by a single principle, the theory of electromagnetism, physics had simpler more elegant theory of the way physical objects behave (Smolin, 2007). In their quest for simplicity, power, and coherence, some physicists research ways that all physical phenomena could be explained by a single principle. String theory is a candidate for such a task (Smolin, 2007). These physicists can be characterized as looking for a unifying principle to explain the physical realm. Likewise, Kagan can be characterized as looking for a unifying principle “to explain the moral realm” (Kagan, 1989, p. 13).
The unifying principle that Kagan gives can be stated as: ‘One should always act in ways which can be reasonably expected to lead to the best consequences overall.’ By circumscribing morality with this principle. Kagan is able to give a theory that is powerfully coherent and simple. With this principle one is able to evaluate the morality of every act. Potentially one can show whether one should download music or if a group is being wrongfully and systematically disadvantaged.
However, as Kagan convincingly argues, the implication of this principle is that most of our moral beliefs are false. What we are left with is a powerful simple and coherent theory, but not a theory which describes moral phenomena. For by basing morality on a principle, Kagan denies out only access to the moral realm, our intuition. Kagan sacrifices accuracy for elegance.
Kagan’s argument is elucidatory in that it shows clearly how principled based moral theories detach from moral intuition and why this is problematic. However, most principled based moral theories do not stray so far from common moral belief. Indeed much of the persuasiveness of some forms of deontology and consequentialism is that they often match up nicely with intuition. Kagan’s argument is important because it shows that a moral theory based on principles does not necessarily explain moral phenomena. So long as a moral theory is based on a principle it is not built on the data of morality. Even if principles are cooked up in such a way so that the moral values they give fit moral intuition exactly or if the principles themselves are inspired by intuition, the connection between theory and phenomena would still only be contingent. We would have no way of knowing if the connection would continue to hold or not.
Toward an Unprincipled Account of Morality
We should therefore resist the temptation to give unifying principles of ethics. Instead we should embrace unprincipled moral theories. Principled morality fails not because the true general principles of morality haven’t been discovered yet but because there are no general principles of morality. The fact that morality is so resistant to generalization should give us good reason to believe that morality is not a category of unified phenomena. Morality is after all, our own concept. Michael Williams (2001) makes the same point when speaking about epistemology.
When we ask ‘what is knowledge?’ in a philosophical tone of voice, we are not trying to get a grip on the natural world which, in some sense, exists independently of us. Rather we are looking for a reflective understanding of evaluative practices that are our own creation (Williams, 2001 p. 14).
We could just as easily substitute ‘epistemology’ for ‘morality’ in Williams’s statement. It is a mistake to imagine a realm of moral facts out there independent of us. Morality has at its origin in our conceptualization. There is no reason to assume that what we call morality must be unified in any consistent way any more than any other of our concepts.
In fact, Gerald Dworkin (1995) argues in his paper “Unprincipled Ethics” that psychological research supports the idea of an unprincipled morality. He cites work done by Coleman and Kay on lying. In their 1981 study. Subjects were given stories wherein a character was dishonest. They asked the subjects in each case to judge whether or not the particular utterance was a lie. They started with the idea that a lie is as a falsehood that id deliberate and intended to deceive. The stories given t subjects had cases of every possible combination of these three elements. Interestingly, they found that no one combination of these characteristics was necessarily judged a lie (Coleman & Kay, 1981). Dworkin argues that these results do not support a model of moral reasoning where moral values are derived from general principles. Rather, Dworkin claims the lesson from this study and psychological research in general, is that there is evidence that determines the way we reach judgments and solve problems in many areas is not though deductive reasoning from general rules. He argues that moral reasoning is no different (Dworkin 1995, p. 237). There need be no general principles of morality for us to reason morally.
Rather than by a general principle the category of morality can be understood as internally related by what Wittgenstein calls family resemblances (Wittgenstein, 2001). To explain, imagine a family portrait. All the members of the family may resemble each other, although there may be no one characteristic that they all possess. If we wanted to give a general rule by which to describe the similarity of their appearance we could not. We can see the similarity, we can give examples of similarities, but we can make no sharp boundaries. Likewise, we are able to vaguely describe what a certain right action has in common with another. For example, they may both contribute to the general good or both fulfill some obligation. But we could not say that it is by virtue of this characteristic that the action is good. In short we can make no general rule about what constitutes good conduct, although what counts as good conduct is generally clear to us.
Just as principled accounts of morality sacrifice descriptive accuracy for the sake of instructional power, an unprincipled account of morality seems to be in danger of sacrificing the ability to give instruction for the sake of descriptive accuracy. The appeal of principled morality is that it can tell an agent what is right and wrong in virtually all cases. Without principles it is unclear how an account of morality is able to give insight in cases where an agent is unsure how to act, or even unaware that they should be acting in a certain way.
To understand how it is possible for an unprincipled account of morality to give instruction beyond that of moral intuition, it helps to understand how it is that principled ethics is able to do so. The unifying principles given by the moral theories we have discussed act as functions. In order to determine whether an act is right or wrong an agent simply plugs the hypothetical act into the function and the function gives a moral value. For instance, a consequentialist can decide if lying is wrong by asking if lying contributes to the general good. A deontologist can decide if lying is wrong by asking whether they could will that lying become a universal law. In both cases the morality of the act is determined by the principle. In this way, we see that both theories attempt to close morality under the laws of deduction. Potentially any act can be morally evaluated. An unprincipled account of morality obviously can’t make such an appeal to principles. However, acts can be evaluated in more ways than running them through a deductive formula.
Jonathan Dancy (2007) argues that it is possible for an unprincipled account of morality to give insight into how we should act. To do this we must give up on the idea of deriving moral imperatives from general principles. In other words we need to accept that morally cannot be closed under deduction. Dancy explains that in order to get an unprincipled project off the ground we will have to work out a conception of moral judgement not as function of a given principle, but as a “sensitivity to the nature of the situation we find ourselves in and to the demands that it places on us” (Dandy, 2007 p. 2). Dancy argues that every situation has a structure, “a structure of reasons, features that combine with each other to make it the case that here we should o this instead of that” (Dancy, 2007 p. 3). According to Dancy the competent moral judge is the one who is able to recognize the morally relevant features of a situation. Dancy explains that we should not ask “in what way is my decision here determined by previous decisions, or a general principle?” But rather the question should always be “what is the nature of the case before us?” (Dancy, 2007 p. 3)
By focusing on the demands of each situation one is able to decide what features are morally relevant. It may be that in some situations consideration of the general good may seem important, or that compliance with duty is relevant, or perhaps one should consider some other factor. The point is that moral relevance is context sensitive. This flexibility allows morality to be linked to intuition in a real way. At the same time unprincipled accounts do not leave us without means of evaluating moral statements, just without deductive means. We are still able to make specific arguments about what acts are right or wrong by pointing to particulars of a situation. By looking at a given situation, an agent is able to decide the morality of an act when their intuition is silent. Also, by looking at a given situation an agent is able to be innovative in their moral expression. It is possible to make arguments with reference to particulars of the circumstances that common moral practice is inadequate and needs revision. Although unprincipled approaches to ethics do not rest on a foundation of general rules, they are still able to account for our moral practice. We need only to overcome the vertigo of morality without a foundation of general principles.
In summary, by examining Kagan’s argument about the implication of consequentialism I have uncovered a problem that faces all accounts that unify morality by a single or set of principles. By basing morality on principles, such theories divorce theory from the actual experience of morality. Any similarities that exist are only continent. It is clear then that coming to agreement about the underlying principles of good human conduct is impossible. Nevertheless, that doesn’t mean we are without means of learning and teaching what it means to be good. Rather, being good is not a single property but a group f diverse characteristics held together like the overlapping strands of fibre in a rope (Wittgenstein, 2001). How one should act will depend on the context. Therefore, in an effort to promote good conduct we should not concern ourselves with reaching principles and the deductive laws associated with them. Instead, we should be teaching a keen sensitivity to the nature of our circumstances.
Coleman, L., and P. Kay. (1981). Prototype Semantics: The English Verb ‘lie’, Language, 57, 26-44.
Dancy, Jonathan. (2007) ‘An Unprincipled Morality’, in r. Shafer-Landau (ed.). Ethical Theory: an Anthology, (pp 771-4). Malden, MA: Blackwell.
Dworkin, Gerald. (1995). Unprincipled Ethics. Midwest studies in philosophy, 20, 224-28.
Fricker, Miranda. (2006). Powerlessness and Social Interpretation. Episteme: A Journal of Social Epistemology, 3, (1-2), 96-108.
Kagan, Shelly. (1989). Against ordinary morality. Gloutenshire; Clarendon Press.
Kant, Immanuel. Fundamental principles of the metaphysics of morals. (T.K. Abbot trans.). New York: Dover. (original work published in 1785).
Mill, J.S. (1863). Utilitarianism. London: Parker, Son and Bourn.
Ross, W.D. (1930). The good and the right. Oxford: Oxford University Press.
Wittgenstein, Ludwig. (2001). Philosophical investigations. Malden, MA: Wiley-Blakcwell.